Welcome to Geog 486! In this lesson, we will talk about the basics of map design, including how to customize your map to fit a specific audience, medium, and purpose. We will also introduce some topics that we will cover more in-depth later in the course, including visual variables, scale, and online map distribution. For this week’s lab activity, we will be making general-purpose basemaps in ArcGIS Pro. For those of you that haven’t used ArcGIS Pro before, this will be a good introduction to the software, and for all it will provide an opportunity to start thinking more deeply about the principles of cartographic design.
Throughout the lesson content, you will notice Student Reflection prompts. These prompts are opportunities for you to pause and reflect on what you have learned and how it relates to previous course content or your own personal or professional experience. Though not required, you are welcome to post responses to these prompts in the lesson discussion forum. You should post something to the lesson discussion each week, but you may choose to post a question/answer or comment about the lab instead.
Now, let's begin Lesson 1.
Action |
Assignment | Directions |
---|---|---|
To Read |
In addition to reading all of the required materials here on the course website, before you begin working through this lesson, please read the following required readings:
Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allows. |
The required reading material is available in the Lesson 1 module. |
To Do |
|
|
If you have questions, please feel free to post them to the Lesson 1 Discussion forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
“Not only is it easy to lie with maps, it’s essential.” – Mark Monmonier, How to Lie with Maps, pg. 1
When making a map, it is impossible to map everything. In fact, to be a useful model of our world and of any phenomena in it, maps must always obscure, simplify, and/or embellish reality. These actions—which make maps useful—also make their construction subjective. Cartographic design, even when informed by well-established conventions, is an art as much as a science. Every design choice a cartographer makes ultimately influences the map readers’ comprehension, appreciation—and even trust—of the map that he or she creates.
Though maps may include or be supplemented by text or other media (even by sound, smell, or touch), map creation at its core is about visual design. As such, cartographers often talk of graphicacy and its importance in facilitating visual communication with maps (e.g., Field 2018, pg. 194). Graphicacy was first defined by Balchin and Coleman (1966) as “the intellectual skill necessary for the communication of relationships which cannot be successfully communicated by words or mathematical notation alone.” Graphicacy—like literacy—has its own grammar and syntax, and learning the rules of graphic language is essential for designing effective maps (Field 2018, pg. 194).
Which map of the two below best communicates the trend of the data? Why?
In the map on the left (Figure 1.1.1), the rainbow color scheme makes it easy to view the states as grouped into categories by hue, but the lack of an obvious order between the selected colors makes the overall trend unclear. A sequential color scheme (right), however, makes it easy to view the trend of the data, as low-to-high values as are encoded intuitively from light to dark.
The design decisions that go into making a map often go far beyond choosing a color scheme for a simple state-by-state choropleth map. The map below is a Russian Civil War map – flames and smoke are used as symbols of the Bolshevik uprising. This map not only communicates information; it conveys emotion.
As demonstrated by the examples above, the way in which you design a map can deeply influence how your readers interpret it. A well-designed map can intrigue and even surprise its readers, leaving a meaningful and memorable impression. Shown below is a map of projected future storm surge in New York City, designed by Penn State alum and cartographer Carolyn Fish. The map doesn't ask the reader to imagine what NYC might look like under future climate scenarios - it shows them.
Following cartographic conventions—such as applying sequential color schemes for sequential data—typically results in more effective maps. However, some maps diverge from these guidelines. Learning cartographic best practices will help you to both apply them—and thoughtfully disobey them—when prudent.
View the maps in Figures 1.1.4 through 1.1.7 below: Do you think they are effective? Is there anything you think should have been done differently?
Maps are generally classified into one of three categories: (1) general purpose, (2) thematic, and (3) cartometric maps.
General Purpose Maps are often also called basemaps or reference maps. They display natural and man-made features of general interest, and are intended for widespread public use (Dent, Torguson, and Hodler 2009).
Thematic Maps are sometimes also called special purpose, single topic, or statistical maps. They highlight features, data, or concepts, and these data may be qualitative, quantitative, or both. Thematic maps can be further divided into two main categories: qualitative and quantitative. Qualitative thematic maps show the spatial extent of categorical, or nominal, data (e.g., soil type, land cover, political districts). Quantitative thematic maps, conversely, demonstrate the spatial patterns of numerical data (e.g., income, age, population).
Cartometric Maps are a more specialized type of map and are designed for making accurate measurements. Cartometrics, or cartometric analysis, refers to mathematical operations such as counting, measuring, and estimating—thus, cartometric maps are maps which are optimized for these purposes (Muehrcke, Muehrcke, and Kimerling 2001). Examples include aeronautical and nautical navigational charts—used for routing over land or sea—and USGS topographic maps, which are often used for tasks requiring accurate distance calculations, such as surveying, hiking, and resource management.
In theory, these map categories are distinct, and it can be helpful to understand them as such. However, few maps fit cleanly into one of these categories—most maps in the real world are really hybrid general purpose/thematic maps.
Advancements in technology and in the availability of data have resulted in the proliferation of many diverse types of maps. Some, as shown in Figure 1.2.5, are embedded into exploratory tools intended to inform researchers and policy-makers.
Other maps are intended for a wider audience but share the goal of uncovering and visualizing interesting relationships in spatial data (Figure 1.2.6).
Maps also are not limited to depicting outdoor landscapes. Some maps, such as the one in Figure 1.2.7, are designed to help people navigate complicated indoor spaces, such as malls, airports, hotels, and hospitals.
For a map to be useful, it is not always necessary that they realistically portray the geography they represent. This map of the public transit system in Boston, MA (Figure 1.2.8) drastically simplifies the geography of the area to create a map that is more useful for travelers than it would be if it were entirely spatially accurate.
Maps that show general spatial relationships but not geography are often called diagrammatic maps, or spatializations. Spatializations are often significantly more abstract than public transit maps; the term refers to any visualization in which abstract information is converted into a visual-spatial framework (Slocum et al 2009).
Though there are many different types of maps, they share the goal of demonstrating complex spatial information in a clear and useful way. Rather than attempt to place maps into discrete categories, it is generally more productive to see them as individual entities designed to suit a particular audience, medium, and purpose. We will discuss this more in the next section.
Though you won’t need to understand the biology of the human brain and visual system, making great maps requires understanding how people perceive visual information. When discussing how people interpret maps, we can frame this discussion in terms of perception, cognition, and behavior.
Perception in map design refers to the reader’s immediate response to map symbology (e.g., instant recognition that symbols are different hues) (Slocum et al. 2009).
Cognition occurs when map readers incorporate that perception into conscious thought, and thus combine it with their own knowledge (Slocum et al. 2009). For example, readers might be able to interpret a weather radar map without its legend due to their previous experience with a similar map, or might incorporate knowledge of a map’s topic into their interpretation of a visual data distribution (e.g., the higher concentration of people aged 65+ shown in some Florida cities makes sense given what I know about retirement communities).
Behavior refers to actions that go beyond just thinking about maps. Considering how design may influence behavior is essential in anticipating the real-world effects your maps may have. The way a map is designed can influence its readers’ actions and decision-making, and these decisions may range from small (e.g., for how many seconds will the reader look at this map?) to great (e.g., will this flood-risk map convince the reader to purchase insurance?).
Another useful way to think about map communication is with the cartography-cubed model (MacEachren 1994). The model MacEachren (1994) proposed focuses on how different maps and visualizations are used. Within this framework, any map can be located within the cube by determining its location along three dimensions: (1) from public to private (with regards to the map audience), (2) from presenting knowns to revealing unknowns (e.g., is the map for displaying known information or for exploration?), and (3) from low to high interaction (e.g., a static map vs. an exploratory interactive mapping tool).
These dimensions are often correlated, hence the shown corner-to-corner continuum from visualization to communication. A printed map in a magazine article, for example, we could classify as a tool for communication, while an exploratory mapping tool designed for epidemiologists would be better described as (geo)visualization.
Return to the previous section (Types of Maps). Where would you place each of the maps shown within the cartography cube?
There is no inherently good map—only a map that is well-designed and properly suited to its audience, medium, and purpose. Before creating a map, you should ask yourself (and if possible, your clients) several questions (Brewer 2015).
In this course and beyond, you will make many different kinds of maps. Some will be advertisements, some will be scientific documents - some may be just for fun. No matter the mapping project or process you use, pausing to reflect upon the who, what, and why of your map will always lead to better results.
Consider a mobile or desktop mapping application that you use frequently, such as Google Maps. What changes might you make to this mapping tool if a client asked you to alter it for a different, singular purpose—for example, as a wayfinding tool for young children, or for assisting police during emergency response?
Chapter 1: Planning Maps. Brewer, Cynthia. 2015. Designing Better Maps: A Guide for GIS Users. 2nd ed. Esri Press. (note: pages. 1-3 are required reading for this lesson, but you may find the rest of the chapter helpful as well.)
Basemaps are essential – they provide the context for your map data. Selecting a basemap should never be just an afterthought, and though the final choice is always subjective, you can make a better decision by considering your map purpose, audience, and the nature of your overlay data.
Often the default basemap used in business web-mapping applications. Helpful when highly-detailed locational context is necessary (particularly for navigation). Though pre-designed street basemaps may not have the ideal aesthetic for overlaying complex data, they are particularly useful at large scales (at which they appear less visually cluttered) or when overlaying relatively simple social data (e.g., for a map showing all locations of a restaurant chain).
Often useful for environmental or engineering applications. May be useful in rural areas that cannot be well-understood using street maps (as few streets exist). The colors and detail make overlay data much more challenging to design than over subtle basemaps – satellite basemaps work best when GIS data is structured and simple and understanding the physical structure of the landscape is essential to the mapping function (e.g., for a map of local water pipelines).
Usually reserved for thematic mapping, greyscale basemaps are helpful when the intended audience already knows the location context, or when significant detail is not important to fulfill the map’s purpose. The simple backdrop adds visual emphasis to your overlay data – especially important for maps produced for entertainment or maps whose primary focus is statistical data (e.g., statistical mortality maps). Choose a light or dark background based on the content and mood of your map, and design overlay data accordingly.
Terrain basemaps are particularly useful when the terrain of the landscape has an important relationship with the data being mapped (e.g., mapping wildfires; hiking maps). Shaded relief also adds visual interest and, when done well, creates a beautiful map. Just be sure to not let the basemap content overwhelm your own data.
A comparison of several example basemaps at the same location in Chicago are shown below in Figure 1.5.5. As shown, different basemaps can have vastly different overall looks, as well as differing levels of detail (LOD).
When making a map, your basemap sets the tone - everything else builds from this important beginning.
There are many more options for basemaps than the defaults available in ArcGIS Pro, though they are a great place to start. Have you used any mapping applications that you felt had an exceptionally-designed basemap?
Check out some more creative, exciting basemaps in the Mapbox Gallery! [20]
There are many creative possibilities - visit Mapbox's selection of Designer Maps [21].
Though many pre-designed options exist, and can be selected as described above, the best reference map for a specific task is often the one you make yourself. When downloading base data for a map, you should consider the following data layers, of which you might need a few or many. This is not an exhaustive list of available base data content, but will help you start thinking about the kinds of data you may need.
A good basemap will often include data that shows the shape of the physical landscape. All terrain layers are typically derived from a digital elevation model (DEM), which is a grid-based (raster) data layer that contains elevation layers.
Elevation can be mapped in several different ways; a common method is hypsometric tinting (hypso) or coloring based on elevation values, shown in Figure 1.6.1.
Contour lines are often used to show more detail about the shape of the landscape, either alone or combined with hypsometric tinting, as shown below.
Other layers such as hillshade and curvature are often added for additional visual detail.
Orthoimages, or images of the earth’s surface that have been properly transformed for mapping purposes, can also be used alone or combined with terrain layers. We'll talk more about terrain visualization later in the course.
Political boundaries are often important components of basemap design. Commonly-mapped boundaries include international borders, state or province boundaries, incorporated places, smaller census units such as tracts and blocks, and boundaries of Native American reservations, among others. Place names are used to add additional locational context.
Other layers that can be useful as base data include zoning and land use data. These data are often available in vector form from local GIS organizations. Land cover and impervious surface data, among other layers, are available in raster form from the National Land Cover Database (NLCD).
Hydrography can also play an important role in a basemap. Data used may include streams, rivers, lakes, swamps, marshes, and wetlands, among other water features.
Given the vast amount of data available, it is important to think carefully about the base data necessary for map’s audience, medium, and purpose—and design accordingly.
Chapter 2: Basemap Basics. Brewer, Cynthia A. 2015. Designing Better Maps: A Guide for GIS Users. Second Edition. Redlands: Esri Press.
When designing your maps, two ideas should be at the forefront of your symbol design process: (1) order, and (2) category. Map symbol design relies heavily on the proper use of visual variables—graphic marks that are used to symbolize data (White, 2017).
Cartographer Jacques Bertin (1967) was the first to present this system of encoding data via graphic elements. Suggestions of supplemental visual variables (e.g., transparency), as well as analyses of their utility in different cartographic contexts, have been brought forth by multiple well-known cartographers (e.g., MacEachren 1994).
Some visual variables (e.g., size, color saturation, and color lightness) clearly indicate quantitative changes in magnitude. These are best for encoding data that has an order (e.g., a county-level map of population density; a road map with both highways and local roads). Other visual variables (e.g., color hue, pattern, and shape) signify qualitative—but not quantitative—differences. These are best applied when data categories have no inherent ordering (also often called nominal, or qualitative data), such as in a choropleth map showing political boundaries.
Figures 1.7.2, 1.7.3, and 1.7.4 demonstrate how visual variables can be used to symbolize common features in general purpose maps. These variables can be used either independently or in combination, to create the best visual representation of the underlying data.
Edward Tufte, a statistician and data visualization expert, said “the commonality between science and art is in trying to see profoundly—to develop strategies of seeing and showing” (Zachry and Thralls 2004, pg. 450). The goal of cartography, both an art and a science, is to optimally visualize—and help others see—the world, and various phenomena within it. To do so takes patience, practice, and skill—all of which you will continually develop throughout this course.
Do a simple web search for maps of a topic that interests you. What visual variables are used in these maps? Are they effective?
Another important decision you will have to make when mapping is at what scale your map should be designed. When designing your symbols, you should always take scale into consideration. Generally, large-scale (zoomed-in) maps should include more features, such as local roads and points of interest, while small-scale maps should be simpler, to avoid visual clutter.
What do you see at the four different scales shown in Figure 1.8.1? What features are prominent at the smallest scale (top left)? What features do not appear until the largest scale (bottom right?)
Web-based basemaps, such as the one shown in Figure 1.8.1, are often designed to adjust the level of detail automatically, as the user adjusts the map’s scale. If you are mapping your own data over a web map, however, you will still need to make decisions about the level of detail you include at each scale, as well as the sizes and styles of your symbol designs.
Cynthia A. Brewer & Barbara P. Buttenfield (2007) Framing Guidelines for Multi-Scale Map Design Using Databases at Multiple Resolutions, Cartography and Geographic Information Science, 34:1, 3-15, DOI: 10.1559/152304007780279078
Particularly given the rise of web-based mapping, maps are readily shared across the web – occasionally even going viral. Many blogs and websites are host to many maps, though the most efficient way to share maps with many users is through social media. Social media platforms like Twitter are an easy way to share both interactive and static maps, as well as links to external sites that permit more than a 140-character explanation of your work.
Maps can serve many purposes – from communicating ideas to exploring data to generate new insights. For any purpose, however, maps must be read and used. Social media has its faults, but it is an excellent way to get design inspiration and to share your own work.
Have you ever used Twitter or other social media sites to share or learn about maps? It can be a fun way to stay up-to-date on current mapping trends, and to gain and share new ideas. Cartographer Kenneth Field shared this list of cartographers and (geo)visualization experts on Twitter in his recent book, Cartography. (Field 2018). If you’re interested in learning more about (the most) current trends in cartography, you may find it a helpful place to start. ESRI.com Cartography Sources and Resources [28]
During this course, we will be completing some peer critiques. However, for your first map critique, you will be critiquing a map made not by one of your peers, but by an external source. As noted by cartographer Kenneth Field (2018), the ability to thoughtfully reflect on the design of a map is an important skill. It will improve both your own map-making abilities, and your ability to comprehend the maps of others.
To complete this assignment, you should write-up a 500 word (max) critique of one of the following maps:
Map #1: Where to Live to Avoid a Natural Disaster - NY Times [29]
Map #2: Monsters of the Mekong - National Geographic [30]
Map #3: Grand Canyon Panorama - National Parks Service [31]
In your written critique please describe:
As suggested by the prompts above, map critique is not just about finding problems, but about reflecting on a map overall. Your critique should focus on things the map does well as much as it does on suggestions for improvement. In your discussion, you should connect your ideas back to what we have learned in Lesson One. You are also welcome - but not required - to relate the map to a personal or professional project or experience.
Please list the title of the map you have chosen at the top of the page.
A rubric is posted for your review.
This week, we'll be making two (2) general-purpose 8.5" x 11" maps. In addition to being an introduction to map-making in ArcGIS Pro, this lab brings together a variety of concepts discussed in this lesson. When making these maps, you'll need to consider
- scale
- visual variables
- map symbols
- audience, medium, and purpose.
All the requirements for this lab are listed below: you should reference this page as you work, and before you submit your final maps.
For Lab 1, you will create two (2) general-purpose maps in ArcGIS Pro.
There is a rubric for your review.
More instructions are provided in Lesson 1 Lab Visual Guide.
Before you look through the Visual Guide, please watch the following video (6:28) entitled "Lesson 1 Lab ArcGIS Pro Tips & Tricks." Doing so will give you a few hints on how to start with this lesson. You should not expect to follow the video exactly as the map design process and the decisions made on how to design the map is up to you.
(0:01)
This is the ArcGIS Pro starting screen which you'll see when you open the map file for Lab 1.
Here are some of the base maps we read about in Lesson 1. I've pre-selected to include the gray canvas basemap which we'll be working with for this lab. The basemap also comes with a reference layer (that you can use to help locate an area of interest to map) that you can toggle on and off, but we won't be including any labels on our map in Lab 1. I encourage you to toggle off the basemap before you submit the map as the basemap includes labels that disturb your design. Besides, we will work with labeling in the next lesson.
For Lab 1 and 2, we'll be working in Louisiana. The data we'll be using was all downloaded from The National Map, and I've pre-loaded all the features you will need. You can expand groups of layers as well as toggle layers on and off in the contents pane, which is on the lefthand side of the ArcPro environment. All the data you'll need for this lab is in the database. You won't really need to worry about this for Lab 1 unless you accidentally delete a layer from your map. In that case, you can drag it back onto the map from the database. For example, if we accidentally remove the roads layer we could drag it back.
(1:10)
When designing symbols, it's often helpful to focus on one layer at a time, so I'm going to toggle all but this rails layer off. We can right-click on the layer and then select the symbology option to open the symbology pane. For this layer, we have just a single symbol, which we can edit in the symbol properties pane. Within this pane, the first tab is most helpful for making simple adjustments such as changing symbol color or width. For example, we can change these lines to red and we can increase their width - and you'll see that preview appear at the bottom of the pane.
(2:00)
More detailed edits can be made in the layers and structures pane. For example, we can add an additional layer to create a multi-layer line, and we could also rearrange these layers if we wanted to. Back in the layers tab, we can change the line's colors and make additional edits. It's a good idea to explore all the design options available in the layers tab.
You may notice that our lines have a strange "caterpillar" look. This can be corrected by enabling symbol layer drawing which will fix the ordering of your layers and clean these lines right up. Some layers, such as roads, contain multiple feature types. The roads in this map are classified by their TNMFRC value. Different values signify different types of roads. You can explore this more by opening the attribute table for this layer.
(3:17)
There's some interesting design you can do with area features as well. We'll work from the symbology pane to edit our water bodies; just as we did with lines, we can change the fill color - so let's go into the color picker and do yellow (you wouldn't do yellow, but let's try yellow) and then we can add another fill layer on top. Go back to the layers pane, and we can change this to a hatched fill. I really don't like the yellow let's change to green - so there you have a pretty easy example of creating a pattern effect.
(3:58)
Another helpful feature is the show count option which displays the count of each feature type in your dataset. You can see there's only one feature in this underground conduit classification - and we're just going to remove that. Unlike the codes, which are linked back to the database, you're free to change the labels as much as you want. We'll talk about this more in Lab 2. You can also change the ordering of features by clicking one and using the arrows to move it up or down.
(4:28)
To make the second map for this lab, we'll start by saving our first map as a map file. You should name it something that makes sense and something that you'll remember. Essentially what we're doing is making a copy of our map that we can then re-import into the same project. So let's do that now by choosing the import map file option. Our map isn't done, but imagine it is - so let's try a new layout using the import layout option. Name your layout something that makes sense, and then you're ready to add your map! I'm going to put in some half-inch margins here and then go to the map frame drop down and click on the appropriate map. You can resize and rescale your map once you add it to the page. To change the location and view on your map though, you'll have to activate it. Once activated you can move your map around as much as you'd like. The final step is to add your name and export your map. Onelast step is to use a text box to add my name. Now, Go to the Share tab and export your map as a PDF: we'll increase to 300 dpi.
To start this lab, you'll want to download the zipped folder and copy it to a safe space on your computer that has plenty of file space. I recommend dedicating a folder on your computer or a large external drive just to Geog 486 lab projects to keep yourself organized. There will be several large files used in this class.
To open the starting map file, you'll need to extract the folder and open the blue ArcGIS Pro file called "Lab1_START." It should look similar to the file in Figure 1.1 below.
All the features/data you will need have been downloaded by your instructor and pre-loaded into this project file.
You can toggle on and off layers using the associated checkboxes in the Contents pane. The light-gray canvas basemap has been included as part of these files. The basemap also comes with a reference layer (that you can use to help locate an area of interest to map) that you can toggle on and off, but we won't be including any labels on our map in Lab 1. We will work with labeling in Lab 2. While you can use the World Light Gray Reference and World Light Gray Canvas Base layers as guides during the map design process, make sure that you toggle off both the World Light Gray Reference and World Light Gray Canvas Base layers before you submit your final maps for this lesson.
There are Layers Groups (e.g., “Transportation”) as well as individual layers (e.g., “Roads”). Eventually, you will need to look at multiple layers at once, so that you can see how all your symbols look together. It will likely be easiest at first, however, to turn off (un-check) most of the layers so you can focus on one layer at a time.
The data you see in the contents pane are stored in the project's geodatabase. You can see this data by expanding the database in the Catalog pane. For this lab, you don't have to worry much about managing the data in the geodatabase - the data you need has already been added to your map. If you accidentally delete a layer from your map, however, you can drag it back onto the map from here.
As a suggestoin, it may make sense if you started from the "bottom" layer and worked your way "up." In other words, think about "visually" what is the lowest layer in the list of data. For example, let's assume the area of interest you selected is near a large water body. What color would you assign to that water body? Figure 1.4 shows how to select a layer (here, railroads) and open the Symbology pane. Clicking on the symbol will let you edit its properties. The next layer to work with may be "land." Again, what color do you imagine appropriate for land given you color choice for the water body. How does the land color you selected contrast/compliment with the water color you chose? Upon inspection, you will likely have to change the colors associated with one or more of the layers until you have achieved a visual agreement with all of the layers, their colors, line thickness, and line styles. Continue adding additional layers according to your visual hierarchy.
Looking in the Gallery of the Symbology pane will give you some ideas, but you should alter these symbols - do not accept the defaults.
Design changes (e.g., color; thickness, style) are made in the symbol properties tab (Figure 1.6; left tab of the Symbology pane). Note that for Map 1 in this lesson you must work only in greyscale. Think about symbol ordering/importance as you design - more important features should have greater visual emphasis. Most detailed work is done in the symbol layers tab (Figure 1.6; middle tab). Experiment with the many options available (e.g., offsets and dashes). You can also preview your symbol at the bottom of the pane. The Symbol Structure tab (Figure 1.6; right tab) allows you to make multilayer lines. You can also drag to re-order these lines.
You may notice a strange “caterpillar” effect when you create multi-layer lines. This is due to the default layering of line segments in ArcGIS Pro, but it's easy to fix.
You can fix this layering issue by enabling Symbol layer drawing within that layer from the Symbology Pane.
Some layers, such as roads, have multiple feature types within them - these feature types are specified within that feature's attribute table. For this lab, these have already been classified for you in the Symbology Pane – TNMFRC values are used to specify road types, and FTypes are used for specifying types of waterbodies. Classifying these layers lets us symbolize features based on a crucial attribute, such as road type (e.g., we can make more important road types such as highways more visually prominent).
Similar symbol options are available for area features – for these you will be choosing fills and outline colors/patterns. Experiment with different patterns but be careful with their implementation as patterns can look harsh and visually disruptive: remember that your main map must be designed in greyscale. Exploring the Gallery tab may help you develop ideas.
You are free to alter the labels for each feature type, or change their order using the arrows in the Symbology pane. Note that it doesn't really matter what your labels are for this lab, as long as you understand them. We will not be creating a legend in Lab 1, so these labels will only be visible to you.
You can also drag to re-arrange entire layers within the Contents pane. Think carefully about the ordering of the features on your map. Should railroads be drawn above or below lakes and rivers? What about political boundaries? Why? You may want to reference popular general purpose maps such as Google maps to compare your choices, but there is not always a right answer. Think of your audience and map purpose!
Once you’re happy with your large-scale (1:24,000) map, save it as a map file by right-clicking on the map name in the Contents pane - you should save it in the same folder as this ArcGIS project folder to keep everything organized and connected.
You can then import that saved map into this map project. Note that a map project can contain several different maps and map layouts. Once you re-import your map, this will create a duplicate map within the project file. You can then use this as a starting map for making your smaller scale map. Your main tasks then will be to add color and adjust your symbols for this smaller scale.
Creating a duplicate map this way is not required. Another option is to start your second map from scratch. I recommend creating and editing a copy of your first map instead, as this map will likely have a similar design to your first map, and creating a copy will prevent you from having to re-do a significant amount of design work (unless your second map has a different scope and purpose than the first map).
Staying organized will help you tremendously in the long run. A big part of this is saving your map files with useful file names. Use the Properties dialog box to change your map names to something memorable and descriptive - you don't want to mix them up.
Some ideas for descriptive map names are shown below:
Use the Insert tab to create an 8.5" by 11" layout. Either Portrait or Landscape layouts are fine—but either way, use guides to create a ½ inch margin all around. Once you've created a layout, you can import your map as shown below. Use the labeled map rather than the "default" map to insert your map at the appropriate scale.
Once you've added your map to a layout, you'll want to make some final adjustments.
- You'll need to activate your map as shown below to pan around the area.
- Make sure you've chosen an area of interest that suits the map requirements. It's ok to adjust your map's location at the end - when you designed your map symbols, they were automatically applied to the entire dataset.
- Whether or not your map is activated, you can adjust its scale at the bottom of the page.
- Make sure that you toggle off both the World Light Gray Reference and World Light Gray Canvas Base layers before you submit your final maps. Except for your name, there shouldn't be any labels or text on the map.
- Note that the map in Visual Guide Figure 1.18 is not well-designed at all - it's intended only as an example of how to insert and activate a map in a layout.
Visual Guide Figure 1.18. Activating a map and changing the scale.
The final step is to export your maps as PDFs. Remember you will have two layouts, one for each map. Use the Share tab to export your layouts.
Considerations when exporting. For most maps, a 300dpi is fine. However, if you use
- gradient area fills
- complex area patterns
- coastline effects
then, change the resolution to 150dpi. Otherwise, the file sizes will become extremely large and Canvas can't display these large file sizes. Once your PDF is exported, check the file size. You should keep your exported PDF's file size to less than 10MB. When I go to look at your maps, Canvas has a difficult time displaying fIles larger than 10MB.
Use “Show count” to view how many of each feature type are included in the map data.
Remember to experiment with multiple layers, verify your map design meets all requirements, and design your 1:24,000 map in only greyscale and your 1:100,000 using color. Designing a map in greyscale may require you to be a bit creative with multilayer symbols and patterns - but that's a good thing! As shown in the example below, you can use different shades of grey and patterns or other fill ideas to create interesting map symbols.
That's it! If you have any questions, please post them to the Lab 1 discussion board. You are also encouraged to browse the discussion board if you do not have a question - you may be able to help out a classmate, and you may learn something from a question that someone else has asked.
Credit for all screenshots is to Cary Anderson, Penn State University; Data Source: The National Map.
Now that you’ve finished this lesson, you should have a solid understanding of the importance of visual design, and the many factors that must be considered when making a map. During this lesson, we discussed the importance of considering a map’s audience, medium, and purpose – three vital factors to consider when planning a map.
We also introduced the idea of symbol design, and how to leverage order and category of visual variables to create a more informative map. At the end of the lesson, we touched on issues of scale and map-sharing, which we explore in more depth later this semester. In this lesson’s lab, we began applying this knowledge by building general-purpose maps using ArcGIS Pro and a popular source of open-source geospatial data: The National Map.
You have reached the end of Lesson 1! Double-check the to-do list on the Lesson 1 Overview page [46] to make sure you have completed all of the activities listed there before you begin Lesson 2.
Welcome to Lesson 2! In the previous lesson, we learned the basics of map and map symbol design, and created some general purpose maps in ArcGIS Pro. This week, we're going to focus on what we left out of those maps - most notably, place labels and marginal map elements (e.g., scale bars, north arrows, etc.). We'll discuss typography and the art of text-based elements: you'll learn how to classify and select appropriate fonts, and how to apply this knowledge when creating place labels for maps. Then, we'll focus on another important topic in cartography: the design of a map layout. You'll build and customize a map legend, and practice designing with appropriate visual hierarchy and balanced negative space.
In this week's lab, we'll be working from the maps we designed last lesson. That way, you'll be able to focus on applying the new topics we have learned, rather than starting from the beginning. By the end of this lesson, you will have learned how to create a complete, well-designed general purpose map from open source data. In addition to that being an achievement in itself, these general skills will prepare you for creating more specific, topic-driven thematic maps in labs to come.
Action |
Assignment | Directions |
---|---|---|
To Read |
In addition to reading all of the required materials here on the course website, before you begin working through this lesson, please read the following required readings:
Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allows. |
The required reading material is available in the Lesson 2 module. |
To Do |
|
|
If you have questions, please feel free to post them to Lesson 2 Discussion Forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
When you think of maps, you likely don’t think much about text. In Lesson One, we defined graphicacy—the skill needed to interpret that which cannot be communicated by text or numbers alone—as distinct from literacy (Balchin and Coleman 1966). Despite this, map graphics are often augmented with text, either on the map itself (as in map labels), or in the margins (titles, legends, etc.) Thus, text plays an important role in map design.
View the map in Figure 2.1.1 below—can you immediately tell what is missing? Can you still recognize the location?
As shown above, good label design often employs different colors, font styles, sizing, and more. Map labels play an important role in mapping—not only by labeling symbols, but also by serving as symbols themselves. In this lesson, we’ll learn about the many design effects that can be used to make appropriate text symbols and aesthetically pleasing designs.
Text on maps, as seen in Figure 2.1.1 above, often refers to place names. The study of geographic names is its own subject of study. A commission within the International Cartographic Association (ICA) is dedicated to toponomy, or the study of the use, history, and meaning of place names. If this interests you, you can learn more about toponomy and the ICA on the ICA website [47].
Particularly in thematic mapping, text is employed not just to identify places, but to explain data. In Figure 2.1.3 below, text is used in the making of map legends, scale bars, and so on. Despite this map’s careful color and layout design, without text—it would be unusable.
Place naming is often a contentious and complicated task. Can you think of a place that is referred to differently by those who live there than by those who do not? How do these different names influence the identity of this place?
Rose-Redwood, Reuben S. 2008. “From Number to Name: Symbolic Capital, Places of Memory and the Politics of Street Renaming in New York City.” Social and Cultural Geography 9 (4): 431–452. doi:10.1080/14649360802032702.
“The choices of fonts for uses can be seen as related to the personality of the fonts. The Script/Funny fonts scored high on Youthful, Casual, Attractive, and Elegant traits which are all related to Children’s Documents and artistic elements. The Serif and Sans Serif fonts were seen as more stable, practical, mature, and formal; the uses they are appropriate for fit these characteristics.” (Shaikh, Chaparro, and Fox 2006)
“Make it easy to read.” – Roger Black
There are many elements to consider when designing text for maps. As a cartographer, you want your text to be clearly legible against the map background, be appropriate for the features you are labeling, and match the overall aesthetics of your map.
As you start designing labels, it is best to learn a bit about typographic design.
A typeface is a design applied to text that gives letters a certain style. An example of a typeface is Arial. Many typefaces contain multiple fonts, so typefaces are sometimes called font families. For example, the Arial font family contains several fonts, including Arial Black and Arial Narrow (Silverant 2016). Though it is technically incorrect to do so, the words typeface and font are often used interchangeably. It is less important to understand this nuance than to understand how to apply fonts in practice.
Fonts can be classified in several ways. For example, as text fonts vs. display fonts (Figure 2.2.1).
Text fonts are designed to be simple and legible: examples include Arial, Calibri, Cambria, and Tahoma. Display fonts are decorative fonts like Stencil, Curlz MT, Bauhaus 93, and Castellar. These fonts are often used in branding and for advertisements. Use these fonts with caution, and sparingly on maps. They are perhaps appropriate for a map title, but for little else (Brewer 2015).
Possibly the most common way to classify fonts is as serif or sans-serif (Figure 2.2.2). Serifs are small strokes added to the end of some letters in a font, such as in the widely-recognized font Times New Roman. Sans-serif fonts do not contain these small strokes. Sans-serif fonts as sometimes viewed as informal, modern, and best suited to digital formats; serifs are often described as best for formal print production. These general guidelines, however, are less important than the specific context in which you use a font. In map design, pairing a serif and a sans-serif together in a map often works best.
Though the presence or absence of serifs may be one of the most obvious characteristics of a font, there are many design factors that influence a font's style. Figure 2.2.3 below illustrates many of the different components of type design. Changes to these elements create the difference between different font styles.
Browse the web—or your closet—looking for logos and similar advertisements that employ text as part of their branding design. How does the style of a font change your perception of that brand or item? Do you notice any that work particularly well? Why is this?
There are a wide number of web resources available for learning more about typography—some are linked in the recommended reading section of this lesson topic. Much of this advice, developed for graphic designers, journalists, and others, will also apply to text design for maps. In Designing Better Maps, Cynthia Brewer (2015) outlines several features of fonts that make them particularly useful for cartographers. You should keep these in mind when selecting fonts for your maps.
As shown in Figure 2.2.4, some typefaces contain many font variations. This can be very useful for map labeling, as it permits the cartographer to create distinct labels for different types of features while maintaining a consistent look and feel throughout the map.
You are likely quite familiar with the use of bolding and/or italics to create distinct font styles. A distinction of note, however, is shown in Figure 2.2.5—the difference between an italic and bold font, and bold and italics as applied afterword by a word processing program such as Microsoft Word. Though applied italics and bolding (Figure 2.2.5; right) will work in a pinch, bold and italic fonts designed as a separate font style (Figure 2.2.5; left) take specific characteristics of the typeface into careful account when applying these styles, typically resulting in improved aesthetics and legibility.
Unlike when writing a paper, where most of your text is horizontal and of similar size, the variability of text sizes and angles on a map presents and additional challenge to cartographers. As you will likely use a font in many different instances on your map, a good font choice is one that remains legible when angled and printed small or viewed from a large distance.
X-height has a simple definition – the height of a lowercase x.
A small x-height results in greater distinction between different letters, which is helpful when reading a block of text. When creating labels for maps, however, a large x-height is typically preferred, it results in fonts that are easier to read when printed small on a page.
This one is self-explanatory, though it may not always be possible (e.g., when using most sans serif fonts). Legibility is improved when the reader can tell immediately whether a letter is an uppercase i, lowercase L, or a number 1. The same goes for distinguishing between a zero and an uppercase O. Though typically a zero is shown as a thinner ellipsoid, in some fonts this difference is more distinct than in others.
In addition to selecting proper fonts, there are many design details that can be applied to improve your map labels. These include text color, halos, and shadows, as well as changes to character spacing and sizing.
A halo is often helpful, particularly against busy backgrounds, for helping text display over the background of a map. Halos are distinct from outlines, as they are placed behind text—and they are typically a better choice for legibility, as they do not interfere at all with the text itself (Figure 2.2.8).
Halos are not always as pronounced as the one shown in Figure 2.2.8. Choosing a halo that blends in with the background color of the map creates a subtle look that doesn’t call attention to the halo, but still sets the text legibly apart from any lines that may cross beneath it. See Figure 2.2.9 below – a subdued yellow-green halo blends into most of the background but prevents contour lines from obscuring the legibility of the interval numbers.
Many text effects are available in ArcGIS, and in graphic design software such as Adobe Illustrator. Experiment with text effects when designing your maps, and don’t be afraid to move beyond default settings to create more engaging, legible, and attractive maps.
We learned about visual variables in Lesson One and applied those ideas to create general purpose maps. For example, you might have used different line weights to create hierarchies of road features, or different hues and/or patterns to differentiate between types of waterbodies. In this lesson, we apply these same ideas to text.
Look at the labels on the map in Figure 2.3.1. Which show categorical differences from others? Which show order differences? Which show both?
When designing labels to show order (e.g., population size, (road) speed limit), choose text characteristics that demonstrate differing levels of importance, such as those shown in Figure 2.3.2.
When designing labels to demonstrate category, choose text characteristics that demonstrate difference, but not importance or order (Figure 2.3.3).
As with symbol design, it may often be prudent to use both types of characteristics together—creating labels that show both order and category. When designing labels, be cautious to attend to the aesthetics of your map, and avoid over cluttered or overcomplicated design. It often looks messy to use more than two fonts on a map, so try to stick to two: as noted previously, pairing a serif and a sans-serif font that look good together often does the trick.
Chapter 6: Labels as Symbols. Brewer, Cynthia A. 2015. Designing Better Maps: A Guide for GIS Users. Second. Redlands: Esri Press.
Ideal label placements are always context dependent—many factors, such as the density of map features or character length of place names, will determine the best way to place your labels. Even so, it is helpful to understand best-practice guidelines for placing labels on maps. In this section, we will learn how best to place map labels for point, line, and area (polygon) features. As a cartographer, you will apply these guidelines using both automatic labeling procedures in GIS software and though the manual editing of graphic text.
When placing point labels, two factors are of primary importance: (1) legibility, and (2) association. You don’t want your reader to struggle to read your map labels, and it should always be clear to which point each label refers.
The first guideline to remember is that adding point labels is not like making a bulleted list—your labels should be shifted up or down from their associated point feature. An example ordered ranking of label placements for point feature labels is shown in Figure 2.4.1.
Though the placement ranking guidelines in Figure 2.4.1 provide a good starting point, it is notable that cartographers do not always agree on this specific order. If you are a very astute reader, you may notice that these recommendations vary slightly from the point label placement guidelines given by Field (2018) in this week's required reading. Cartography is not only a science but an art, and sometimes there is more than one right answer. Additionally, while such guidelines are helpful, label placement is a continuous balancing act. Figure 2.4.2 (left) shows two labeled points, both placed at the ideal label position shown in Figure 2.4.1. This arrangement of point labels, however, makes it seem ambiguous to which point “East Gate Shopping Center” refers. In Figure 2.4.2 (right), this label is moved to the second position. The ambiguity disappears.
In addition to the orientation of point labels, you will also need to decide how closely to place them to your point features. In the left image, labels are placed very close to points, while on the right, labels are placed at a greater distance from their associated point symbols. Though map elements that appear too tightly packed are generally undesirable, how closely your labels and points are placed will depend on the size, shape, and density of your labels, points, and map. Most important is maintaining consistency throughout your map design.
Another important consideration is when and where you will apply line breaks to the text on your map. When it fits on the map, showing the entire label on one line (Figure 2.4.3; left) is appropriate. However, due to the density of map features and length of feature names, this is often not possible.
When line breaks are used, place them at natural breaks in the feature name. For example, Mission Hills Country Club looks strange as Mission/Hills Country/Club (Figure 2.4.3; middle) but natural as Mission Hills/Country Club (Figure 2.4.3; right). You should also use spacing between lines that is smaller than the spacing between other labels on the map, clearly demonstrating that these lines of text belong together.
Point labeling is further complicated when labeling multiple types of point features. Your goal should be again to avoid ambiguity—labels should help demonstrate feature categories. As shown in Figure 2.4.4, it is best to label land features on land, and coastal features in water.
Label design is about the details, and often very small changes to label placements can really improve the readability of your map. Figure 2.4.5 below shows how a couple of small edits were used to improve a set map labels. From left to right, line spacing within the “Shawnee Nieman Center” label was decreased to -2 pts., and then the "Nieman Plaza label" was shifted to the left.
Note that though counterintuitive, the use of -2 line spacing, or leading, does not create overlapping lines. Negative leading is generally recommended for multi-line labels—too much space between lines makes them look disjointed, which may cause map readers to incorrectly perceive them as separate labels (referring to separate features).
When labeling line features, similar guidelines as for point labeling exist—design for association, but not at the expense of legibility. Labels should generally follow line features—but not cross over perpendicular lines—as this makes the text harder to read. In some instances, this advice will not be practical, but it is best to first learn the rules so you can more thoughtfully break them.
Figure 2.4.6 shows two maps with labeled streets; the right-sided image is a definite improvement. Unlike in the left map, labels in the right map are aligned with streets and do not cross other lines. Labels in the right map are also better aligned for the eye to understand the naming conventions of the neighborhood: see W 100th Ter, W 101st St, and W 101st Ter, from North to South (maps are North-up). It is much easier to understand this progression in the right map. This sort of line placement is also useful when labeling contour lines, which have an even more important orderly progression.
In lieu of map labels, shields are often used to label highways and other important roads. Though interstate shields in the US are consistent, many states have unique highway shield designs. Using these custom shields in your maps is not always practical, but it can give them local character, and create a better match between the map and the real world.
Similar but additional guidelines exist for labeling non-road line features, such as flowlines. Streams, rivers, and other waterlines should be labeled with text that shows their categorical difference from road features. This is often done with italics (text posture), and/or by using a hue that matches the feature symbol. Figure 2.4.7 shows several examples of labels applied to the stream “Little Cedar Creek”. The label in the map at the left is legible but does not follow the flow of the creek—it looks rigid, as if it is a road label. In the middle map, the label does follow the creek, but this time too much so—it is difficult to read. The label placement in the far-right map is best—a gentle curve makes it clear that this label refers to a water feature, but not at the expense of legibility.
Figure 2.4.9 contains additional examples of line label improvements. Three general guidelines are demonstrated by this figure: (1) follow the feature, but not at the expense of legibility, (2) place labels above lines rather than below, (3) don’t write upside down.
If a line feature is quite long, the label will need to be repeated periodically. The interval at which your line labels repeat is up to you as the map designer and will depend on the map’s feature density, audience, presentation medium, and purpose.
Just as rivers are labeled with curves to follow the flow of water, area features should be labeled in a way that highlights their most characteristic feature: extent. Labels for natural features such as water bodies and mountain ranges should demonstrate their physical extent across the landscape. Use UPPERCASE letters and stretch the label across the area of the feature.
When covering areal extent with labels, focus on finding a balance between character spacing and size. Increasing spacing is generally best—recall that increased font size suggests increased importance. To cover the extent of a feature, however, you may want to increase font sizing somewhat—too distant spacing with a small font size is likely to be challenging to read.
A common mistake to avoid is aligning area labels horizontally across the map frame. Though horizontal alignment is helpful when reading large blocks of text, this design is off-putting when viewed on a map (Figure 2.4.11; top). Stagger area labels for increased legibility (Figure 2.4.11; bottom).
Like regular line feature (e.g., roads, rivers) labeling, avoid labeling across boundary lines when prudent. When labels must cross over map lines, ensure that this does not compromise their legibility, nor overly obscure the feature underneath.
In some instances, particularly for political boundaries, it makes more sense to label the boundary of a feature, rather than its extent. You have likely seen this implemented in maps for navigation, or other interactive basemaps (Figure 2.4.12).
When labeling maps, you will often encounter locations with a lot of features in need of labels; this can pose a significant challenge. Leader lines can be used to connect features with labels that do not fit on or directly adjacent to their respective feature on the map. However, you should not overuse text halos, as these can obscure the map features underneath (Figure 2.4.13; top left). Nor should you overuse leader lines (as shown in Figure 2.4.13; top right)—this leads to a visually confusing map. Instead, find a balance between these techniques; experiment with label hue contrast and use leader lines sparingly. With practice, you will be able to create a well-balanced set of labels, such as shown in Figure 2.4.13 (bottom).
Further improved cartographic design is shown in Figure 2.4.14. This map shows how text color contrast, sizing, and occasional use of leader lines can create a balanced, legible, and aesthetically pleasing map design—even in a complicated map with many labels.
In summary: when creating a map feature label, balance different techniques, and continually ask yourself two over-arching questions: (1) Is the label clearly associated—both in style and positioning—with the feature being labeled? (2) Can I read it?
“Design is as much an act of spacing as it is an act of marking.” – Ellen Lupton, Thinking with Type
When designing a map layout, it is important to design using visual hierarchy—arrangement of graphic elements in a way that signifies their order.
The following map elements are listed by (Slocum et al., 2009) in their general order of importance:
Most of these elements are intuitively named, but the difference between a map’s frameline and neat line can be confusing. A frame line encloses all elements on the map layout, while a neat line crops the map area. A graphic explanation is shown in Figure 2.5.1. A frame line might also act as a neat line if all elements (e.g., legend, title) are shown within the map area (Slocum et al., 2009).
It is typically efficient to place the most important features first, as they will take up the most space on the page. Be cautious, however, not to just start adding items wherever there are holes in the layout—good design is about balancing white space, which does not mean just filling it in. Often, the best way to find a good layout arrangement is to try many different arrangements and note what works. There will never be just one correct way to arrange all map elements.
The graphics and explanations in Designing Better Maps are exceedingly helpful for developing an understanding of layout balance and design. This would be a good time to complete the required reading for this week.
When placing elements on the page, be cautious to leave enough space between them. For example, Figure 2.5.2 below shows how adding just a bit of negative space can result in a cleaner, clearer map design.
Another important component of layout design is the intentional reduction of ambiguity. For example, if your layout includes multiple maps (e.g., a primary and a locator map), and multiple scale bars, it should be clear which scale bar is associated with which map.
Using boxes (e.g., boxed legends) will often seem like an easy solution, but you should use these sparingly, as they tend to create crowding and making aligning map elements more challenging. As you finish designing your layout, ensure that all elements are visually aligned. See the recommended reading below, as well as the required reading for this week, for additional detail and images of proper layout alignment and design.
Chapter 11: Map Elements and Typography. Slocum, Terry A., Robert B. McMaster, Fritz C. Kessler, and Hugh H. Howard. 2009. Thematic Cartography and Geovisualization. Edited by Keith C. Clarke. 3rd ed. Upper Saddle River, NJ: Pearson Prentice Hall.
The part of your map layout that will likely require the most thought—except of course, for your map itself—is your map's legend. A map legend is a key composed of graphics and text that explains the meaning of any non-obvious map symbols. This non-obvious component is important to remember. Consider the general purpose map in Figure 2.6.1 below:
The legend isn’t incorrect, but it doesn’t help explain the map’s already clear design. Did you need a legend to understand that the blue features were water? Every element in your layout takes up precious space—there is no need to waste it explaining symbols that your readers will understand without it.
The same principle applies when adding text to your legend, such as a legend title. Legend titles should be used to add context and explain your map. Don’t title your legend “legend”—your reader will know it is a legend. If there’s no better title then "legend", it doesn’t need a title at all.
If your map does require a legend, use the same care to design it as you do with map symbols and labels. Be cautious of the way you create column breaks or other visual groups in your legend design. People tend to perceive groups of things as related - use this to your advantage in your legend design.
Figures 2.6.2 below shows a choropleth map with an accompanying legend. Though the legend accurately prints the map colors and their matching data values, the splitting of legend items across three columns breaks up the list in a way that may be confusing to the reader.
Below in Figure 2.6.3, the legend design has been much improved. A single column creates an easy visual representation of the color scheme for the reader.
For some legends, you will want not to eliminate column groupings, but to re-position or even create them. In Figure 2.6.4 below, inappropriate column groupings lead to ambiguity regarding the classification of some symbols. Are trails part of transportation, or are they their own category? What about streams? This legend leaves too much up to the reader to interpret.
Figure 2.6.5 shows an improved version of this legend. Note that the different shape of the legend container means that it will need to be placed differently on the page—this highlights the importance of experimenting with layout arrangements throughout the design process.
Note that the examples in 2.6.4 and 2.6.5 contradict the previous statement that obvious symbols like "lake" can be left off the legend. We will slightly relax this "only non-obvious features" guideline in order to practice creating well-designed legends, and due to the presumption that some of our symbol designs may stray far enough from cartographic convention to be nonintuitive to map readers.
Chapter 3: Explaining Maps. Brewer, Cynthia A. 2015. Designing Better Maps: A Guide for GIS Users. Second. Redlands: Esri Press.
In addition to a legend, your maps will often contain other supporting graphic elements such as a scale bar and north arrow. Similar principles apply—you should make your design as simple as possible while still supporting the reader’s understanding of the map. Commercial GIS software such as ArcGIS permits you to easily add accurate scale bars to your map. These will automatically match your map’s scale, and dynamically update if you re-scale your map within your layout. When it comes to visual design, however—be wary of GIS defaults. You will typically have to make manual simplifications to these elements, scale bars in particular.
Figure 2.7.1 shows examples of default scale bar designs inserted into a map layout in ArcGIS, alongside illustrations of their appearance after manual adjustment.
Like making a legend, the first question you should ask yourself before designing a north arrow for your map is: do you need it? Depending on the map projection you use, the direction which points north may not be consistent across your map—in this case, a background grid may be more appropriate. Most maps do use a north arrow, however, and if you do use one, similar conventions to scale bar design exist. Aim to make your design as simple as possible without sacrificing comprehensibility.
View the two scale bars in Figure 2.7.3. In general, as described in Figure 2.7.1, the top scale bar is considered better design. Can you think of a map for which the scale bar at the bottom would be more suitable? Why would it be?
This week, we'll revise and combine our two maps from Lab 1 into one neat, well-designed layout with labels, a legend, and marginal map elements (e.g., scale bars, north arrow). You'll get to build off your hard work from last week and apply new knowledge from this week: typographic design, label symbology, and layout design.
This lab, which you will submit at the end of Lesson 2, will be reviewed/critiqued by one of your classmates as part of Lesson 3. Receiving critique of your work and using this to inform future cartographic design decisions is an important skill to develop. Giving feedback to others also often teaches you new ways of looking at your own and others’ map designs.
For Lab 2, you will create one complete map layout, with a main and a locator map.
A rubric is posted for your review.
More instructions are provided in Lesson 2 Lab Visual Guide.
Lesson 2 Lab Visual Guide Index
Start this lab by opening your project file from Lab 1. Use “Save As” to create a new project for Lab 2. After this, you'll be ready to add labels!
We do not have to write our own labels for map features - they're already in our data - we just need to make them visible. Map features often contain multiple fields (data columns) with possible names, so we need to identify the best ones to use. To do this, open the attribute table for the layer you want to label. We can see the Full_Street_Name field seems like a good option to start with for this layer.
To turn on labels for a layer, right-click the layer and toggle labeling on. To edit the labels, open Labeling Properties as shown below. This will open the Label Class pane. In this pane, the expression box shows how your labels are being drawn from a field in the attribute table. In some cases, ArcGIS will correctly identify the best field to use for labels. In other cases, it will not, and you will have to alter the expression manually. We will use Full_Street_Name, the field we identified earlier.
Begin editing the style of your labels&nbwith the Labeling menu in the ribbon shown below. The default label symbols available are good starting points - they will help give you an idea of how to best design your own labels.
You should also create label classes using this top menu bar. Similar to when we classified roads by their TNMFRC code in Lab 1, we create label classes so that we can create different types of labels within a feature category, and use these classes to design our labels with visual order and/or category.
When you create a label class, all you are creating is a class with a name - ArcGIS will not automatically recognize, for example, that a label class named "Interstates" should only be applied to roads which are interstates. We will tell ArcGIS this using SQL (structured query language).
In our data, all interstates have a TNMRC code of 1 (this code signifies the interstate road-type; see Figure 2.8). We can define this label class using the SQL view in the label class pane. See below:
Note: The Label Class Pane can also be used to create label classes, instead of the top menu bar. You may find it more helpful to use the Label Class Pane for most labeling tasks.
If you forget which TNMFRC code refers to which road type, you should refer to the image below. You can also open this view in your project - your road features should still be classified by TNMFRC code, so viewing it in the symbology pane should create a view similar to the one below.
You should create a different label class for each road type for which you wish to have a different type of label. This includes small differences, such as font size. You do not necessarily have to create a different label class for every road type, but you will likely have several (e.g., local road, collector road, highway, etc.). You should reference the lab requirements page to ensure that you have created enough different label classes throughout your map.
Once you create your label classes, you can switch back and forth between them while editing using the Class dropdown menu. Note that if you create multiple label classes, you will need to define all of them, including the default label class. If you do not, you will have duplicate labels. For example, you may have Interstates labeled in one class, and all roads labeled in the default class - causing interstates to be labeled in both classes.
Another option is to delete the default label class - but be careful when doing so that you are maintaining all the labels you need.
Once you've created a label class, you can use the Symbol tab in the Label Class pane to edit its style. Shown here is the label symbol editing menu (left), and the formatting menu (right). These are used to change many aspects of a label's symbology - including fonts, sizes, spacing, color, etc. Highlighted in green are options I’ve found especially helpful – but don’t limit yourself to these. You should experiment with all options for symbol design—font, weight, spacing, etc. Recall from the lesson content that line spacing (leading) can be a negative value.
In addition to changing the style of your labels, it is important to also assign how they should be positioned. Recall the lesson content on text placement - our goal for this lab is to place labels only with automatic rules. We will not be placing or adjusting labels by hand.
There are many positioning parameters you can adjust in the label position tab—try them out and watch how your labels change. There are a lot of useful options (e.g., Feature Weight) whose function may not be immediately clear to you - I recommend using the link at the end of this sentence to learn more about labeling with the ArcGIS Pro Maplex Label Engine [56].
In addition to simply drawing a label from a feature's attribute table, you can edit label expressions using SQL to append words or other text for more descriptive labels. Don't worry if you haven't done any programming - you only need to make minor edits to create label expressions.
You can also use SQL to append additional text to a label from the attribute table. An example is shown below - though, in this example, you are creating quite a wordy label, which is generally not recommended.
Remember that you will be adding labels both to your large-scale (primary) map, and your small-scale (locator) map. Once you've finished adding labels to your primary map, you can add similar labels to your locator map. You can also save and then import a copy of your large scale map into your project, and then adjust it for the new smaller scale. This is the same process we used to create our second map in Lab 1.
To duplicate/re-import (a refresher):
Your final task is to create a Portrait or Landscape layout with your two maps, a legend, and text elements. An example layout design is shown below.
An example of a portrait layer is shown below: not that your map will also include a title, legend, etc. Additionally, these map examples are not shown in their final form - you are encouraged to use them for layout ideas, but you should not copy their designs.
Before importing your maps, add guides for ½ inch margins – you should not include anything on the page outside of these margins. Note again that the examples below contain unfinished design—they should not be interpreted as examples of finished feature or label symbology.
For the locator map to be useful, you will need to insert an extent indicator. You should do so with the small scale map selected. This will draw a rectangle showing the extent of your large-scale map within the (larger) region covered by your small-scale, inset/locator map.
Marginal elements such as north arrows and scale bars should be added at this point. Keep your North arrow and scale bars simple and easy to read. Use “adjust width” to create clean scale bar values. You can also edit the color, font, label locations, etc., of all marginal elements. Reference lesson content for design ideas.
Another important component of your map layout for this lab will be its legend. Insert a legend with your large-scale map selected so it reflects your large-scale symbol design. Your locator map should use similar symbols, and therefore should not need a legend.
Right-click your new legend element in the contents pane, and choose “Properties” to edit.
You do not have to include every item in your legend, and you may want to change the names of some items significantly. Your goal is to create a comprehensible map. To change the design of different legend elements, select them from the drop-down menu in the Format Legend pane.
You can also make changes from the ribbon.
An efficient way to clean legend titles is to edit the layer titles themselves in the TOC—for example, by opening the properties dialog box for the county boundary layer and changing “GU_CountyOrEquivalent” to “County.”
Once you have made sufficient edits, you may want to disconnect your legend from the data by converting to graphics. This will give you more freedom over the design, but as your legend will no longer update dynamically if you update any map symbols, you should save this step until the end. You will have to “ungroup” the elements to edit them. Once you convert to graphics, you will need to right-click and “ungroup” multiple times to edit the elements for detailed design work. (Note that this is not a well-edited legend, just an example of one in process).
Once your legend is complete, there are only a few final touches to be made. Use the “Dynamic Text” dropdown to move the service layer (basemap) credits out of the map frame and place them elsewhere in your layout, for a cleaner look.
Don't be afraid to re-arrange your layout elements as you go! It may take quite a few tries before you find an optimal design.
Remember to create visual hierarchy for marginalia elements:
Below is an example of a landscape layout made from similar data - you will need to adjust your map to work with the assigned data and location. Note also that the map below may not include all required elements for this lab, but is an example of how your layout might look if you are on the right track.
Credit for all screenshots is to Cary Anderson, Penn State University; Data Source: The National Map.
Here we are - at the end of Lesson 2! In this lesson, we learned about two vitally important but occasionally overlooked aspects of map design: the design of labels and other text elements, and the building of a neat, balanced map layout. In Lesson 1, we discussed visual variables and how they can be used to visually encode order and category in map symbols. In Lesson 2, we extended this idea to include map label design. We also discussed order in another context - the creation of a visual hierarchy in a map layout.
As you likely noticed while working on Lab 2, neither adding labels nor designing a map layout are trivial tasks. Something as simple as creating a legend or scale bar requires significant thought and attention. Little details such as the alignment of layout elements may feel like the "last mile" in the making of a map, but they are key for getting your readers' minds to where they ought to go.
You have reached the end of Lesson 2! Double-check the to-do list on the Lesson 2 Overview page [57] to make sure you have completed all of the activities listed there before you begin Lesson 3.
Welcome to Lesson 3! In previous lessons, we discussed broad concepts related to map and map symbol design, including designing for a map’s audience, medium, and purpose. We learned about visual variables and how to designate order and category with map symbols. In the context of text on maps, we discussed these ideas in greater detail; we created symbols with labels and learned how to place them appropriately on maps. We then put everything together in a map layout.
So far, we have only designed general purpose maps. Though these maps still contain data (e.g., road networks, lakes, boundaries), we have not yet added more abstract data to maps. In this lesson, we discuss thematic maps and the ways in which we can use maps to effectively visualize spatial data. When deciding how to map, we’ll consider the spatial dimensions and models of geographic phenomena, levels of data measurement, and appropriate methods of visual encoding. We’ll compare and contrast the four most common types of thematic maps (choropleth, isopleth, proportional symbol, dot) and map two of these in Lab 3, using a popular data source for thematic mapping – the US Census Bureau.
By the end of this lesson, you should be able to:
Action |
Assignment |
Directions |
---|---|---|
To Read |
In addition to reading all of the required materials here on the course website, before you begin working through this lesson, please read the following required readings:
Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allow. |
The required reading material is available in the Lesson 3 module. |
To Do |
|
|
If you have questions, please feel free to post them to the Lesson 3 Discussion Forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
We first introduced thematic maps in Lesson One, and described them as maps intended to highlight features, data, or concepts (either quantitative or qualitative). In Labs One and Two, we used visual variables to show order and category of typical map features on maps.
The maps we’ve created so far have been general purpose maps—designed to display features of general interest. In designing our maps, we created abstract representations of the real world, with roads, rivers, lakes, county lines, etc., with hues and shapes different from what would be captured by a photograph. Despite this, our designs have still reasonably matched their physical reality. In this lesson, we turn to more abstract depictions of the world, designed using thematic data. View for example, the map in Figure 3.1.1.
This map uses color value—not to show category or hierarchy of map features—but to visually encode county-level unemployment data. Figure 3.1.1 also simplifies the map of the US (not showing even major highways or mountain ranges, but only state and county boundaries) to emphasize the map’s theme.
Due to thoughtful use of color and a simple layout design, this map successfully communicates geographic trends of unemployment in the United States. But was this the best choice to show this concept? Might there be a better way?
View the map in Figure 3.1.2 below.
This map uses a similar color scheme and layout, but encodes its data primarily using proportionally-sized symbols. Color value is used for additional effect, a technique called dual encoding.
These maps (3.1.1; 3.1.2) are both correct—they fit into cartographic conventions. But there are other maps that this cartographer could have made with each dataset that would have been equally correct. There are also maps they could have made that would have been—arguably—quite wrong. How to decide?
Do you think the data mapped in Figure 3.1.1 would be appropriate for making a proportional symbol map (e.g., Figure 3.1.2)? Why or why not?
Before beginning the how of making a map, we need to take a step back and consider the what—the geographic phenomena we want our map to be about.
Geographic phenomena are elements that exist over geographic space. When we say geographic we typically mean anything between the size of a human and the size of the world. So, while still spatial, the connections between the neurons in your brain or the arrangement of atoms in a ceramic material do not constitute geographic phenomena. In this lesson, you will learn tools for conceptualizing, visualizing, and communicating the many phenomena that do.
US Census Bureau. 2021. “Interactive Maps [58].” Accessed May 31.
Geographic phenomena are often classified according to the spatial dimension best used to describe their nature. These include points, lines, areas, and volumes (3D). As you likely remember, we used the spatial dimension of map elements (e.g., line vs. point) in the last lab to decide how to symbolize and apply feature labels to our maps.
Points exist in a singular location and thus have theoretically zero dimensions. Points are usually specified using a coordinate pair (x, y) of latitude and longitude, though they occasionally include a z (height).
Lines are one-dimensional spatial features, typically defined by a series of (x, y) coordinates. A z (height) dimension can also be assigned to lines, but this is uncommon. Lines are used to map phenomena that are best conceived of as linear features, including both some features that have greater dimensionality in reality (e.g., rivers) and those that do not visibly exist in the real world at all (e.g., property lines).
Area features are two dimensional and are represented by a series of (x, y) points that enclose a space. Areal phenomena include natural features like lakes and parks, as well as human-defined locations—from continents to census blocks.
2-½ and 3-D features are sometimes grouped together, but the distinction between them is important. 2-½D features define a continuous surface—they have an x, y, and a z at every location. A good example is elevation, which varies continuously across the landscape. Therefore, a topographic map is a common depiction of 2-½D phenomena.
True 3D maps have an x, y, and z, plus an additional data value, at every location. Imagine, as an example, a map of elevation like the one above; but at every point along the terrain surface, there are additional measurements being taken at various depths. Thus, rather than depicting a continuous surface, true 3D maps depict a continuous volume.
Keep in mind that the scale of your map has significant influence on what spatial dimension will best represent the phenomenon you intend to map. Cities, for example, are usually drawn as areas on large-scale maps, but appear as points on smaller-scale maps. Rivers are usually drawn as lines on small-scale maps but are better represented as areas on large-scale maps. We will discuss this more during discussions of cartographic generalization later in the course.
When conceptualizing the geographic phenomena we want to map, it is important to consider the best way that these phenomena can be modeled. In general, we can categorize the best model for a given phenomenon as existing somewhere along two continuums: (1) from discrete to continuous, and (2) from smooth to abrupt.
You likely learned the difference between discrete (e.g., as shown by a histogram) and continuous (e.g., as shown by the bell curve) variables in an introductory statistics course. The distinction in cartography is similar.
Discrete phenomena have well-defined boundaries: they occur at specific locations, with space in between. Examples include people, cars, houses, hospitals, and roads.
Continuous phenomena, conversely, have ill-defined or irrelevant boundaries. Examples include temperature, air quality, and elevation.
Phenomena can also—independent of their classification as discrete or continuous—be considered either smooth or abrupt.
Smooth phenomena are those that change gradually over geographic space. Examples include precipitation levels and solid aridity: they vary by location but do not typically change abruptly at geographic bounds.
Abrupt phenomena do change suddenly at a geographic boundaries, whether physical or cultural. Often, phenomena are not clearly smooth or abrupt, but fall somewhere in between. The amount of pesticides in soil, for example, might vary somewhat continuously over space, but change rather abruptly at political boundaries (e.g., due to changing government regulations).
Figure 3.3.1 illustrates various surfaces used to represent geographic phenomena throughout the discrete to continuous and abrupt to smooth continuums. Keep this idea of a continuum in mind—geographic phenomena often cannot be classified into neat categories, and it is typically more fruitful to think of them as “more continuous” or “more discrete” than to try and fit them into a box.
Identify the proper (approximate) location in Figure 3.3.1 or the following phenomena: Health insurance (% of people covered); water quality; political affiliation; surface porosity. Why did you place them where you did?
Figure 3.3.2 above shows different map representations that are suited to mapping the geographic phenomena located at these relative positions along the continuous-discrete and abrupt-smooth continua. We will discuss the appropriateness of various thematic mapping methods further later in this lesson.
MacEachren, Alan M. 1992. “Visualizing Uncertain Information.” Cartographic Perspectives 13 (13): 10–19. doi:10.1.1.62.285.
Considering the characteristics of the geographic phenomena you wish to map will inevitably improve the quality of your maps. However, before you design your map, you must understand the distinction between the characteristics of the phenomena and those of your data.
Consider again the map from Figure 3.1.1.
This map illustrates unemployment rates in the US at the county level. Though it is a well-designed and attractive map, consider the characteristics of unemployment as a geographic phenomenon. The abrupt change in unemployment rates at county boundaries in this map obscures the underlying heterogeneity in unemployment within county bounds. The phenomenon of unemployment varies by person, while the mapped unemployment data varies by county. This doesn’t mean the map is wrong, but it is a reality important to be cognizant of, both while creating your own maps and while critiquing those designed by others.
Relatedly, when creating maps, you will often rely on data that has already been collected by others. Often, this data is collected (as in the example in Figure 3.1.1 above) by enumeration units, such as counties, census tracts, or states. Unemployment does vary by person, but it is unlikely that this fine-grained data will be available to you. If you have somewhat coarse (e.g., state level) data, you cannot create a map that shows variation by person, by county, etc., even if this would be a more accurate depiction of the phenomena. The only way to create a more detailed map is to collect more granular data. Your map design can always be altered to present a simplified depiction of your data—but not the other way around.
Slingsby, Aidan, Jason Dykes, and Jo Wood. 2011. “Exploring Uncertainty in Geodemographics with Interactive Graphics.” IEEE Transactions on Visualization and Computer Graphics. doi:10.1109/TVCG.2011.197.
Data is typically classified as either qualitative (e.g., land use; political boundaries) or quantitative (e.g., per capita income; temperature)—you likely recall learning about this distinction in earlier courses. The classification of your data as qualitative or quantitative will have significant influence on which visual variables you use to map your data. Color hue, for example, is excellent for qualitative data, while color value demonstrates order and thus is a better choice for designing quantitative maps.
Nominal is a common term used to described qualitative, or categorical data. Land use and land cover maps are popular examples of nominal data. They might show, for example, residential blocks as distinct from parks and green space, but this does not suggest that one is lesser or greater than the other.
Quantitative data can be further classified as ordinal, interval, or ratio data.
Ordinal data has an order, but cannot be presumed to show differences in magnitude. Sports team rankings, for example, describe which teams are better, but not by how much.
Interval data describes orders of magnitude but has an arbitrary zero point. The classic example is temperature: 30 degrees is warmer than 10 degrees, but it’s not necessarily three times as warm. Another good example is shoe size—you can say that a size 12 is larger than a size 6, and that (unlike if it were ordinal data) there is more difference between a 6 and a 12 than between a 12 and a 13, but the 12 is not twice as large as 6.
Ratio data, conversely, has a non-arbitrary zero point. Examples of ratio data include counts of forest fire incidence, and yearly household income (e.g., $50,000 is twice as much as $25,000). Interval and ratio data are often grouped together and classified as numerical data.
View the map in Figure 3.4.2 above—is the data shown qualitative, ordinal, interval, or ratio? How does this compare to the likely level of measurement of this data when it was first collected?
Consider time—would you usually consider this to be nominal, ordinal, interval, or ratio data? Why?
Chang, Kang-tsung. 1978. “Measurement Scales in Cartography.” The American Cartographer 5 (1): 57–64. doi:10.1559/152304078784023006.
Understanding your data’s spatial dimensions, geographic model, and levels of measurement will help you select which visual variables to use in your map. Recall the table of visual variables we first encountered in Lesson One (Figure 3.5.1). This is a good time to check your own knowledge and consider which of the follow seven variables are best for visualizing data category, and which are best for visualizing order.
Some visual variables are also better than others for encoding data with different levels of measurement. Bertin (1967) only considered size (other than position on the map) to be a truly quantitative variable, its visual representation able to be matched precisely to a numerical value. This makes it a good choice for mapping ratio-level data, as making mathematical calculations with such data can be useful. Visual variables that can typically encode only category, not order (e.g., color hue; shape) are best for qualitative data.
Note that the visual variables presented in Figure 3.5.1 are those originally proposed by Bertin, and though they are arguably the most common still in use, this is not a comprehensive list. The graphic also does not demonstrate the many ways in which these variables might be altered and/or combined to create new designs. At the end of this lesson, we will assess a variety of maps, many of which use multiple visual variables. We will also discuss multivariate mapping further in Lesson 7 (Multivariate and Uncertainty Visualization).
The figures above focus on geometric visual variables (e.g., color; pattern; size), though another common mapping technique is to use pictographic or iconic symbols (Figure 3.5.2).
Iconic symbols are those that provide a closer visual match to their referent, or the real-world element meant to be depicted by the map symbol (Maceachren et al. 2012). The map in Figure 3.5.3 below uses flower symbols that are drawn similarly to how they appear in reality to create an engaging and useful map. It is important to balance usability and realism when using iconic symbols on maps - ensure that they do not become overcrowd, or distract from the map's purpose.
Another important consideration that should be weighed when considering the use of iconic symbols is the cultural context of those symbols. Iconic symbols can have meaning only to a specific group of individuals. For example, imagine that you are driving down the interstate and see a road sign that shows a knife and fork. Some people would understand these icons to imply food. However, the knife and fork icon is not necessarily understood to imply food by individuals who, for example, only use their hands while eating and may have never seen a knife and fork. Iconic symbols, therefore, are very culturally contextualized and that context should be weighed before icon symbols are chosen to be used on a map. Here is an article [66] that further explores the idea of symbols and icons and their meaning in cartography.
Like other continuums we have discussed (e.g., discrete to continuous; abrupt to smooth), map symbols cannot always be classified as simply abstract or iconic, and instead, exist somewhere in the middle. National Geographic's Atlas of Happiness [68] for example, uses smiling face graphics to encode data about happiness. Thus, it is less abstract than if this data had been encoded only with color value or size, but less iconic than if more realistic graphic images of people were used.
Visual variables are used in many mapping techniques: in addition to selecting which visual variables you use for your maps, you will also need to choose what type of thematic map you will create. The four most popular thematic map types are choropleth, isopleth, proportional symbol, and dot maps.
This would be a good point to complete the required reading for this week, particularly pages 81-91 in Thematic Cartography and Geovisualization. The reading gives an excellent overview of visual variables and thematic mapping techniques.
The required reading gives more detailed descriptions, but below we give a general overview of the four most popular types of thematic maps.
Choropleth Maps are maps in which color or shading is applied to distinct enumeration units, usually statistical or administrative areas. Color hue, saturation, and value are the most frequently used visual variables in choropleth mapping, though pattern is sometimes used as well. Choropleth mapping should not be used to encode exact counts (e.g., number of people living in each state), as the visual encoding of color by enumeration units makes this confusing. For example, consider that more people live in California than in any other state. You could create a state-by-state choropleth map showing counts of, say, universities or gas stations, and California would likely lead in both. But a map showing this would not be interesting—California has more people and things because it is a bigger enumeration unit. The map would tell us nothing interesting about California's system of education, or its residents' consumption of gas.
Isopleth Maps are like choropleth maps in that they typically use color value to encode data values, but unlike choropleth maps, they do not visualize the enumeration units from which they are built. Isopleth maps are preferred for mapping phenomena that vary continuously over space, as they better represent the distributions of these phenomena than choropleth maps. The primary disadvantage of isopleth maps is that they require quite a bit of data to design them accurately. They should also not be used to map data that change abruptly at administrative boundaries (e.g., percent sales tax). Choropleth mapping is a simpler and more appropriate method for mapping such data.
Proportional Symbol Maps are best for mapping abrupt, discrete data; they visualize data using the size of a symbol (most often a circle) placed inside an enumeration unit. As the symbols are scaled only based on the data value—irrespective of the size of the enumeration unit—this permits the reader to not only view the variation between symbols, but also perform a visual comparison of the size of the symbol and the size of the enumeration unit over which it is placed. Note that the map in 3.5.6, unlike the previous two maps (3.5.4 and 3.5.5) displays count data (population) rather than a rate (percent in poverty; people per sq. mile). This is an appropriate choice for a proportional symbol map.
Size, the visual variable used in proportional symbol mapping, should not be used to map standardized data such as rates (e.g., people per sq. mile). When mapping count data such as population counts, you should use a proportional symbol map, or you should standardize your data before using it to make a choropleth or isoline map. We will talk more about standardization in Lesson Four.
Dot Maps are like proportional symbol maps in that they are excellent for visualizing discrete data. Rather than displaying a different-sized symbol per enumeration unit, however, dot maps are constructed by filling enumeration units with a count of symbols (usually dots) based on the count of the variable of interest within the unit. Thus, this technique is preferred over proportional symbol mapping for mapping data which vary more continuously over geographic space.
It's important to think critically when creating and reading dot maps. Often, dot maps are made by scattering the appropriate number of dots randomly throughout each enumeration unit. To a novice viewer, they give the illusion of high precision - you might assume that if every dot represents one person, that the dots are placed on the map exactly where those people live! However, this is very rarely the case. We will explore the differences between dot and proportional symbol maps more in the lab at the end of this lesson. As you will see, which method is most appropriate depends not only on what phenomenon you are mapping, but also on the scale at which you map it.
Chapter 5: Principles of Symbolization. Slocum, Terry A., Robert B. McMaster, Fritz C. Kessler, and Hugh H. Howard. 2009. Thematic Cartography and Geovisualization. Edited by Keith C. Clarke. 3rd ed. Upper Saddle River, NJ: Pearson Prentice Hall.
Note: This chapter includes the 10 pages of required reading for this week, but if you have access to the text, you may find the additional pages in the chapter useful as well.
Analyze the maps shown below. For each map, identify the level of measurement of the data mapped. What visual variables are used to encode this data? Is the map effective—does it tell you what you need to know?
During this course, we will be completing several critiques. In Week One, we critiqued a map from an external source. Participating in these critiques will improve both how you think about cartographic design skills and your ability to critically evaluate the map design of others.
For this second critique, you will engage in a peer-to-peer review (or peer review). In this activity, you will be assigned a colleague's map from this class to critique. Your assignment for this peer review includes writing up a 300+ word critique of one of your colleague’s Lesson 2 Lab.
In your written critique please describe:
According to the two prompts above, a map critique is not just about finding problems, but about reflecting on a map in an overall context. Your critique should focus on things the map does well as much as it does on suggestions for improvement. In your discussion, you should connect your ideas back to what we learned in the previous lessons.
You may find these two resources helpful as you write your critiques:
Submission Instructions
You will work on Critique #2 during Lesson 3 and submit it at the end of Lesson 3.
Step 1: When a peer review has been assigned, you will see a notification appear in your Canvas Dashboard To Do sidebar or Activity Stream. You should also receive an email notification. Upon notification of the Peer Review (Critique), go to Lesson 2: Lab 2 Assignment. You will see your assignment to peer review one other classmate.
Step 2: Download/view your classmate's Lab.
Step 3: Write up your critique using the prompts above in a Word document. Be sure to also review the rubric in which you will be graded for Critique #2 for more guidance. Save your Word document as a PDF. Use the naming convention outlined below. Please list the name of the student you have been assigned to critique at the top of the page.
Step 4: In order to complete the Peer Review/Critique, you must
- Add the PDF as an attachement in the comment sidebar in the assignment.
- Include a comment such as "here is my critique" in the comment area.
- PLEASE DO NOT complete the lesson rubric as your review, award points, or grade the map you are critiquing. Even though Canvas asks you to complete the rubric, PLEASE DO NOT COMPLETE THE RUBRIC OR ASSIGN POINTS/GRADE.
Step 5: When you're finished, click the Save Comment button. You may need to refresh your browser to see that you've completed the required steps for the peer review.
Note: Again, you will not submit anything for a letter grade or provide comments in the lesson rubric.
This week, we'll be making two map layouts with the same data, so as to compare two different thematic mapping techniques. During this lab, you'll also be downloading some of the data yourself, and choosing a census variable and location based on your interests. This lab is our first thematic mapping lab, but when mapping, you should integrate your design knowledge from previous lessons, such as techniques for creating balanced map layouts and neat map marginalia. The lab requirements are listed below: you should reference this page periodically as you work on the lab, and review the lab rubric as well before submitting your work.
For Lab 3, you will create and submit two map layouts. One will be composed of three proportional symbol maps; the other will be composed of three dot maps. Your final task will be to write a reflection that compares these two techniques in the context of this lab.
Note: The mapped state must contain at least 30 counties.
The following states meet this requirement: Alabama, Arkansas, California, Colorado, Florida, Georgia, Idaho, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Michigan, Minnesota, Mississippi, Missouri, Montana, Nebraska, New Mexico, New York, North Carolina, North Dakota, Ohio, Oklahoma, Oregon, Pennsylvania, South Carolina, South Dakota, Tennessee, Texas, Virginia, Washington, West Virginia, Wisconsin.
A rubric is posted for your review.
More instructions are available in the Lesson 3 Lab Visual Guide.
This is your starting file in ArcGIS Pro: It contains state and county boundary data for the entire US. Your maps will focus on a state of your choosing from the list given in the lab document.
For each mapping technique (proportional symbol; dot map), you will create three maps: state-level, county-level, and census-tract level. In addition to using the boundary files provided, you will download and add census tract boundaries, and data from the US Census Bureau's American Community Survey.
Visit US Census Bureau: TIGER/Line Shapefiles [81]. Select the appropriate data using the drop-down menus. Download and paste the zipped folder into your Lab 3 folder, extract all files, and import as a feature class into ArcGIS Pro (see Figures 3.2 and 3.3).
Add your newly-imported census tract data to the Tract_Map. At this stage, you should have three maps with boundaries: state, county, and census tract.
You will be downloading your census data using the US Census Bureau Data Tool [82]
The ACS census data that you download will include demographic data of your choosing for the state, county, and tract level geographies. This demographic data will be in spreadsheet format that you will then join to the TIGER boundary files.
If you use the Advanced Search option in the census data tool, you may find it easier to search for Census data according to specific topics, geography, years, surveys, or codes. For example, assume I am interested in choosing ACS 2015 five year surveys for all census tracts in Ohio for the purpose of examining the number of veterans. Here is what I would search on using the Advanced Search option. The words listed below correspond to the search criteria in the Advanced Search interface. Note that you can narrow your serach for data by clicking on any of the search criteria in any order. Here is the order I used.
For each geographic scale (state; county; tract), open the appropriate Excel file from your downloaded data folder and follow these steps:
1) Delete the top row (you only want one header row; this will become your top row/field names in ArcGIS).
2.) Save-As each Excel file (one per geographic scale) as a *.csv formatted file with a sensible name (so you can easily find it to import).
As you scan through your file, you will see that there are likely a lot of data columns. Choose one column (variable) that interests you. Most of these columns (variables) you will not use for this lesson. Therefore, it is also helpful to delete the columns you don't need for your map, as this will make the table easier (faster) to import and deal with in ArcGIS.
Use the import table(s) function to import your CSV files. Hit F5 (refresh) if your data appears to be missing! It likely just needs to be refreshed.
Refresh your database in the catalog pane to see the tables you have imported.
For each map, we want to join our imported ACS table to the spatial boundary file.
To do this, we want to find the field that matches between the ACS table and the spatial attribute table – we will join them using this field. Figure 3.10 below shows that the STATEFP field in the US_States boundary files matches the Id2 field in our Census data table.
We have one problem: the values in the STATEFP field (boundary file) are stored as text values, but those in Id2 are stored as numbers.
The easiest way to remedy this is to create a numerical field in our spatial boundary data attribute table and use that field to join it to the ACS data table.
Choose “Double” as your Data Type. Don't forget to save!
We will calculate this field by requesting that the new values be equivalent to those in the original text-based field (STATEFP) we wanted to use. In essence, we are creating a duplicate field with the same values, except that this time the values will be stored in numerical form.
Your spatial file should now be ready to be joined to your ACS data!
Add the join, using your newly calculated field and the matching field in your ACS data.
A Few Notes on Joining the State, County, and Tract Level Data.
- For the State table: when formatting the state *.csv data in Excel, add an ID column to the right of GeoID and type the value of the last 2 digits in GeoID (after "US") into each field, ex. 0400000US01 should be 1 or 0400000US10 should be 10.
- For the County table: do the same thing as directed above but type all of the digits following "US."
- For the Census Tract table: instead of creating a field in the Excel file, go into ArcGIS Pro and create a Text field within the census shapefile. Next, calculate the new field to equal this expression: "1400000US" + !GEOID!. Then join the Census Tract table to the census shapefile using that new field.
Before symbolizing your maps, repeat the previous steps (create and calculate new field; join ACS data) for your county-level and census tract-level maps. You will then have three maps ready to symbolize.
Use the Proportional Symbols method to symbolize your data.
Create a new standard-size (8.5" x 11") layout as you did in Lab #1. Insert a map frame and copy and paste it in the layout: this will create three maps at the same scale.
This (Figure 3.20) is just a quick layout example and should not be considered finished! You should follow the guidelines we have learned for creating a visual hierarchy and an excellent layout, etc.
Use the “Save-As” function and save a copy of your map project with a new name (like YourName_DotDensityMaps). This way, you won’t have to add any new data. Creating your second layout will be much easier this way - instead of doing data joining, downloading, etc., you can just focus on the design!
Have fun adjusting your symbolization as appropriate! Try out different colors and symbols. Experiment with the parameters to see what works! The best design will depend on many factors, including the scale of your map, your chosen state, and the data you are mapping.
You are free to edit your data in this lab as needed to clean it up. Delete states (or any type of row) as needed. You may want to make any unneeded boundaries invisible instead, as this will make it easier to bring them back if necessary.
You can change the number of legend items using the symbology pane.
Reference current and previous lesson content for design ideas. Test several layout configurations and lots of symbol designs (sizes; colors; outlines; transparency) – you’ll learn a lot as you go!
Credit for all screenshots is to Cary Anderson, Penn State University; Data Source: US Census Bureau.
We've reached the end of Lesson 3! In this lesson, we learned a lot about thematic maps - what they are, why we design them, and how to choose the best thematic mapping technique based on characteristics of your data and of the geographic phenomena you wish to map. We discussed challenges you might encounter when making thematic maps, such as when the level of measurement of the data available to you doesn't match the level of measurement of the phenomena. In Lab 3, we created both proportional symbol and dot density maps - exploring the differences between map types and their appropriateness at different scales.
Though our focus this lesson was on mapping data, you'll notice that concepts we learned earlier - such as visual variables, map labels, and layout design - have remained of high importance. The tasks in this course are intended to build upon each other. I look forward to watching you thoughtfully integrate concepts from throughout the course into your maps each week.
You have reached the end of Lesson 3! Double-check the to-do list on the Lesson 3 Overview page [83] to make sure you have completed all of the activities listed there before you begin Lesson 4.
Welcome to Lesson 4! Last lesson, we learned about thematic maps, including how to choose a thematic mapping method and adjust our designs based on the characteristics of a geographic phenomenon. This week, we focus on a specific type of thematic map: choropleth maps. Choropleth maps are the most popular thematic map type, and designing them properly relies on adequate understanding of other important topics in cartography, such as data standardization and classification methods. Choropleth maps also typically employ color in their design: in Lesson 4, we discuss color in-depth. You will learn about the different ways in which we can model color space, and how visual perception constraints - both in the general population, and in those with color-vision impairments - influence map perception.
In Lab 4, we'll explore how choosing a different color scheme and/or classification method can alter how readers interpret your maps. We’ll also learn how to make maps that work well in pairs—a common task that is often significantly more challenging than making one map that stands alone.
Action |
Assignment |
Directions |
---|---|---|
To Read |
In addition to reading all of the required materials here on the course website, before you begin working through this lesson, please read the following required reading:
Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allow. |
The required reading material is available in the Lesson 4 module. |
To Do |
|
|
If you have questions, please feel free to post them to the Lesson 4 Discussion Forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
Color is frequently used to symbolize information on maps. In recent years, cartographers have begun to employ color more and more: in a study of map-color use in scientific journals, White et al., (2017) found that the use of color in published map figures increased from 18.4% in 2004 to 69.9% in 2013. This trend can primarily be attributed to the loosening of practical map production constraints. The cost of color printing, for example, is no longer prohibitory. This is in large part due to the increasing popularity of web-based dissemination of maps and other visual graphics, which makes such costs irrelevant. Tools such as ColorBrewer and Colorgorical have also made color selection easier; the first of these is even now integrated into the color selection tools in ArcGIS Pro.
In this lesson, we will explore the basics of specifying, mixing, and selecting colors for maps. You should aim to understand and properly apply the color schemes available in GIS software, as well as to alter them as appropriate based on your maps’ audience, medium, and purpose. Eventually, you might even design your own color schemes from scratch.
You might remember the map in Figure 4.1.2 from Lesson 1. This map is a thematic map, and more specifically, a choropleth map. Discussions of color in mapping often focus on choropleth maps. This is for good reason—choropleth mapping is the most common thematic mapping technique, and its employment typically requires thoughtful analytical use of color. We will discuss the details of choropleth mapping later in this lesson, but note that color is frequently used on other types of maps as well. General purpose maps often employ color to delineate between kinds of features, and maps that focus on other visual variables (e.g., proportional symbol maps) often also use color to encode an additional variable, or to add visual interest.
When you hear the word "color," words such as blue, red, and green likely spring to mind. Though these are colors in the colloquial sense, these are better described as color hues. When using color as a visual variable, each color is specified not just by color hue but by three dimensions: hue, lightness, and saturation (Figure 4.2.1).
Color is produced when light is either reflected off of (e.g., a car; a printed map) or emitted by (e.g., a computer screen) an object. Hue refers to the wavelength of that light, from longest (oranges and reds), to shortest (blues and violets). Figure 4.2.2 shows nine swatches of color with different hues, in the order of the rainbow spectrum.
In mapping contexts, hue is typically used to differentiate between features. In general purpose maps, for example, hue is used to create categories, and to help the reader identify different features as belonging to a particular group. In Figure 4.2.3, for example, color is used well, and improves the legibility and aesthetics of the map. Though multiple types of roads are shown, all roads are shown in red. Similarly, all hydrologic features and labels are shown in blue - a familiar color easily recognizable by map-readers as associated with water.
Lightness is another dimension of color; it describes how perceptually close a color appears to a pure white object. Lightness is also commonly called value, though cartographers sometimes avoid that term, as value is also used to describe data values—using the same word for both items can cause confusion. Lightness works well for visually encoding the order and/or magnitude of thematic data values—typically, lighter colors signify lower data values (i.e., less signifies less), and darker, more visually-prominent features signify higher data values.
The third dimension of color is saturation. Saturation is also sometimes called chroma. In map design, saturation is generally less important than hue and value, but it still can play an important role. Highly saturated colors are particularly useful for calling attention to small but important map elements that would otherwise be lost (Figure 4.2.4). Caution should be used when using saturation in this way, however—the use of too highly saturated colors, particularly over large areas, may be distracting or accidentally overemphasize those features.
These three dimensions (hue; lightness; saturation) were originally identified by Dr. Albert H. Munsell in the early 20th century. Munsell’s first color model, a color sphere, was an attempt to fit these three dimensions of color into a regular shape. Though this model was still a breakthrough, Munsell realized that it was quite insufficient, as human color perception is not linear and cannot be accurately modeled by a regular shape. The final shape he landed on looks more like a lopsided ellipsoid.
Figure 4.2.5 below takes a top-down approach to visualizing this color space: each of the four graphics demonstrate what is, in essence, a slice of the Munsell model, with increasing lightness from left to right. As shown, the colors which the human eye can perceive do not change linearly through color space—this makes color specification and design a challenging task.
Imagine you want to create a categorical map with a large variety of colors. What does Munsell’s model suggest about what kind of colors would be best used for this purpose?
Though Munsell’s model is helpful for understanding color perception, and perhaps for sharing color specifications with others, a working knowledge of other models is required for building color schemes in GIS and graphic design software. When specifying colors, it is important to consider the display medium that you are using to create them. When mixing paint, cyan, magenta, and yellow (CMY) are used. As mixing paint (or laser printing inks) results in less light being reflected from the color surface, this is called subtractive mixing. The opposite occurs on digital display screens, which create colors by mixing red, blue, and green (RGB) light. Mixing these primaries is called additive mixing.
ArcGIS offers a wide selection of color model choices for specifying colors, including RGB, HSV, and CMYK. RGB and CMYK color models refer to the aforementioned models for mixing additive and subtractive primaries, respectively. RGB is useful for digital media, and CMYK is the color language typically used by graphic artists. Another popular model is HSV (hue, saturation, and value). HSV is reminiscent of the Munsell model (see Figure 4.2.8), but with much greater symmetry—recall the oddly-shaped structure of Munsell’s model.
The symmetry of HSV makes it fit much better into the language of computers, but as human color perception is not linear (recall Figure 4.2.5), using HSV can cause problems unless you remain cognizant of this shortcoming.
Additional color models, including HSL and CIELAB, offer other ways of specifying colors. We will not go further into the details of color specification here, but you are encouraged to explore the recommended readings for more information.
Chapter 7: Color Basics. Brewer, Cynthia A. 2015. Designing Better Maps: A Guide for GIS Users. Second. Redlands: Esri Press.
When applying color schemes to maps, there are many factors to consider. First and foremost, keep this rule in mind: the perceptual structure of the color scheme should match the perceptual structure of the data. For example, if your data go from high to low (sequential data), you should use a color scheme that demonstrates this order, as shown in the map in Figure 4.3.1.
There are three main types of color schemes: sequential, diverging, and qualitative. A popular tool for choosing color schemes on maps is ColorBrewer, designed by Dr. Cynthia A. Brewer at Penn State. ColorBrewer’s interface is shown in Figure 4.3.2. You may find it helpful to explore the many color schemes available on the site as you read more about types of color schemes in this lesson and consider how you might apply them to your maps.
Sequential color schemes are the most popular color schemes used in thematic mapping, as they are excellent for demonstrating the order of data values. Several examples of sequential color schemes are shown in Figure 4.3.3.
Though color lightness is effective on its own, sequential color schemes are also often designed with multiple harmonious hues, such as in the color schemes shown in Figure 4.3.4. The multi-hued nature of these color schemes can make it easier for viewers to discriminate between all data classes on the map. They also often create more aesthetically-pleasing visualizations. As long as it doesn't take away from readers' comprehension of your data, why not make a better-looking map?
As shown in Figure 4.3.5, when hue is paired with lightness it can create a dramatic effect in a sequential map. When making such maps, ensure that they accurately reflect the progression of your data—it is challenging to create an effective sequential color scheme that relies heavily on hue.
Diverging color schemes are similar to sequential color schemes, as they also demonstrate order. Instead of showing a single progression, however, they visualize the distance of all values from a critical point. These color schemes work well for depicting data that have a critical middle value or class (such as maps showing percent change).
If your data has a natural midpoint—such as the absence of change— a diverging color scheme works well, as it permits the reader to easily identify values on the map as either above or below that value. An example of this is shown in Figure 4.3.7 below.
Other values can also serve as helpful midpoints in mapped data. For example, a map might use a diverging color scheme to demonstrate values that fall above or below the data’s mean, or perhaps some external goal-worthy value (e.g., a choropleth map of median income where a diverging color scheme is centered around a calculated value of a living wage).
An important consideration when applying a diverging color scheme is whether your data has a critical break or a critical class (Figure 4.3.8). Using a diverging scheme with a critical class will highlight a critical group of areas on your map, as well as those above and below. A critical break will show all areas as either above or below a critical value –there is no “neutral” color class in this scheme. Diverging schemes also do not always have to be symmetrical. Your critical break/class will often be near the center of your data range, but it in no way needs to be.
Keep the divergent schemes shown in Figure 4.3.8 below in mind as we discuss data classification for choropleth mapping later in the lesson.
View the map in Figure 4.3.9 below. Why is a diverging color scheme used here? What does the map tell you? What doesn’t it tell you? Would you design it differently?
The third type of color scheme is the qualitative color scheme. These schemes are used to demonstrate differences—but not order—between map features. Several examples are shown in Figure 4.3.10 below.
Qualitative color schemes are often used when creating maps of political boundaries, or to create categorical choropleth maps, such as the one in Figure 4.3.11. As the term choropleth is composed of the Greek words for “area/region” and “multitude,” it is technically incorrect to refer to a map of nominal values as a choropleth map, despite the characteristic enumeration-unit shading such maps employ. These maps should instead be called chorochromatic maps.
Perhaps the most common use of qualitative color schemes in mapping is in land use/land cover maps. These maps seek to demonstrate category (e.g., residential vs. commercial) but not to demonstrate order. An example of a land cover map is shown in Figure 4.3.12.
The (color vision unimpaired) human eye can discriminate between about twelve different hues; dependent on the reader and the design of the map, often even less. Many maps, and land use/land cover maps in particular, contain more than this number of categories. A frequent strategy is to group categories into hue classes (e.g., green for vegetation) and then to use lightness and saturation to create intra-class differences. In Figure 4.3.12, for example, green hue is used for forest, and lightness variations are used to differentiate between forest types.
View the categories of land cover in Figure 4.3.12. Does the perceptual structure of the data match the perceptual structure of the colors assigned? Does it do so in more ways than one?
So far this lesson, we have talked about multiple ways to specify colors, and how we might apply them to maps. As we discuss color, however, we also need to discuss color vision deficiency—the inability to discriminate between certain (or occasionally, all) colors. Though color blindness varies by gender and ethnicity, you can generally expect that about five percent of your map readers will have some form of color deficiency. You may have some form of color vision deficiency yourself.
The good news is that several web tools exist to help you design more accessible maps. Viz Palette, developed by Elijah Meeks and Susie Lu, is one useful example. It permits you to import your own color schemes from popular color-picking tools such as ColorBrewer and view their appearance through the eyes of those with different types of color vision deficiencies.
Tools such as Viz Palette are useful for understanding how different people might view your data visualizations and maps. You can then decide for yourself whether your chosen palette is acceptable. ColorBrewer also lets you select from among only color schemes that have been empirically-verified as colorblind friendly - its interface includes an option to show only “colorblind safe” color schemes. Unsurprisingly, the scheme in Figure 4.4.1(2) does not appear.
How much you factor color accessibility into your map design will depend greatly on its audience, medium, and purpose. Color discriminability is affected by many factors outside of genetics, including reader age, lighting conditions, and map resolution. It is also more crucial in some mapping contexts than in others. A map for entertainment, for example, may sacrifice accessibility for increased aesthetics and visual interest among the not color-vision impaired. When a map’s purpose is emergency management or vehicle routing, however, the cartographer may place a greater value on ensuring readability for all map users.
Even among those without color vision impairments, human color perception does not come without flaws. View the squares labeled A and B in Figure 4.4.3—do they look the same to you?
Your eyes are deceiving you—these two squares are exactly the same shade of grey. (If you don't believe it, check out the interactive version of this graphic at illusionsindex.org [100]). This is the result of a principle of color interpretation called simultaneous contrast, or induction—colors appear differently, dependent on the backdrop against which they appear.
View the maps in Figure 4.4.4: which colors in the second map (1, 2, 3, 4) do you think match the colors in areas A and B?
Student Reflection answer: The color in A matches the color in 4; the color in area B matches area 2. Is this what you were expecting?
To date, little empirical research in cartography has evaluated the influence of induction on map interpretation, and, thus, few suggestions exist for minimizing its effects in practice. You should, however, anticipate the effects that varied backgrounds will have on the interpretation of your map symbol colors, particularly for maps in which such comparisons are common and/or critical.
So far in this lesson, many of our examples have been choropleth maps—the most common thematic mapping technique, and one which typically makes extensive use of color as a visual variable. In the next section, we will focus on other aspects of choropleth mapping, including data standardization and classification, as a deeper understanding of how these maps are built using data is required for selecting an effective color scheme.
As discussed in Lesson 3, the choropleth mapping technique should be used on standardized data such as rates and percentages—rather than on totals or counts—which are better represented by point symbol maps.
Your data will sometimes be delivered in the proper standardized format. For example, you might have for each enumeration unit in your data a rate, density, or index value. All of these are appropriate for choropleth mapping. Oftentimes, however, you will need to calculate these values yourself. Data from the US Census, for example, is often delivered as count data by enumeration unit but includes a population field which can be used for standardization.
Using the example data in Figure 4.5.1 above, imagine we wanted to map the number of people in each county who are under 18 years old AND have one type of health insurance coverage (Column F). And imagine we created a county-level choropleth map using those Column F values. What would this map tell us? It might tell us a little something about geographic health insurance trends in North Carolina, but mostly it would just show us in which counties more people live.
Remember the importance of map purpose: rather than just making a population map, we want to understand the geography of health insurance coverage for young people. For this, we need to map standardized values. To do so, we can divide the number of under 18-year-olds with one type of health insurance (Column F) by the appropriate universe: the count of items (here, people) that could possibly fall into this category. Since our data value of interest only applies to a specific age group, our universe, in this case, is not all people (Column D), but all people under 18 (Column E).
Some texts and software programs, including ArcGIS, call this process normalization rather than standardization. As suggested by (Slocum et al. 2009) we use the term standardization, as normalization has a more specific meaning in statistics with which we do not want this process to be confused.
Let’s return again to a map that should be becoming familiar, posted now as Figure 4.6.1. Median income is visually encoded in each state as belonging to one of four classes: (1) less than $45,000; (2) $45,000 to $49,999; (3) $50,000 to $59,999, and (4) $60,0000 and more. How were these classes chosen?
One side-step before we discuss data classification: think back to our discussion of types of color schemes—can you think of another type of color scheme that would be effective in Figure 4.6.1? Do you think it would be better?
When the map in Figure 4.6.1 was being designed, the aforementioned classes had to be decided upon – and there are many different ways in which class breaks in median income could have been drawn. So, how do you choose? Rather than simply choosing the default classification scheme that your GIS software suggests, you should think critically about how your data classes are defined. The first decision you should make, however, is not how, but whether to class your data.
Figure 4.6.2 shows an example of two maps—one unclassed (sometimes called a "class-less" map), and one classed. Unclassed maps encode color (usually with lightness) based on the specific value within each enumeration unit, rather than based on a pre-defined class within which the data value falls. These maps are useful as—if designed properly—they may more accurately reflect nuances in the distribution of the data. However, they should not be considered an easy solution to the problem of data classification. They have their own disadvantages, for example, they make it challenging for the reader to match the value encoded in an enumeration unit to its location on the legend.
Before modern GIS software, unclassed maps were quite difficult to create, but new technology has made their design quite simple. Unclassed maps show a more “direct” visualization of the data, while classifying maps gives you more control over the final map. It will be up to you as the map designer to decide whether to class your map; however, many map readers—and cartographers—still prefer classed maps.
As you will likely be classifying your maps, it is important to understand how this process can influence your final map design. Most of the commonly-used classification methods are available in ArcGIS, and the software interface gives a good simple explanation of each of these methods (Figure 4.6.3). We will not discuss the mathematical details of each of these classification methods here—it is recommended that you explore the recommended readings or do your own research on the web to learn more.
Natural Breaks (Jenks): Numerical values of ranked data are examined to account for non-uniform distributions, giving an unequal class width with varying frequency of observations per class.
Quantile: Distributes the observations equally across the class interval, giving unequal class width but the same frequency of observations per class.
Equal Interval: The data range of each class is held constant, giving an equal class width with varying frequency of observations per class.
Defined Interval: Specify an interval size to define equal class widths with varying frequency of observations per class.
Manual Interval: Create class breaks manually or modify one of the present classification methods appropriate for your data.
Geometric Interval: Mathematically defined class widths based on a geometric series, giving an approximately equal class width and consistent frequency of observations per class.
Standard Deviation: For normally distributed data, class widths are defined using standard deviations from the mean of the data array, giving an equal class width and varying frequency of observations per class.
Though Figure 4.6.3 gives helpful descriptions of each classification method, it offers little advice as to when to use them. A good way to approach this question is to view your data along the number line. You can use histograms (for large data sets) or dot plots (for small data sets) to visualize how your data is distributed, and to select class breaks accordingly. The following suggestions are given by Penn State cartographer Dr. Cynthia Brewer.
1. For data with near-normal distributions, consider classifying your data based on the mean and standard deviation.
2. For skewed distributions, consider systematically increasing classes, such as arithmetic and geometric classing methods.
3. If your data are evenly distributed, equal interval and quantile classing methods work well. These methods are also best for ranked data.
4. Natural breaks, created using Jenks classing method or in selecting breaks by eye, work best for data which shows obvious groupings through the range. The natural breaks method highlights the natural sets of values in the data.
We will look at data using dot plots during this lab associated with this lesson. When you make maps, unless you are working with a very large data set, this will often be the most effective way to visually investigate your dataset in order to choose a classification method or visually/manually place your own breaks. ArcGIS, however, creates histograms of your data that you can also use to understand how the breaks you have chosen relate to the spread of your data.
Compare the breaks, histograms, and maps in Figure 4.6.4 below. Which classification method would you have chosen? Why?
Note that the spread of your data is only one of multiple elements you should consider when choosing how to classify your data. As with other map design choices, your map's intended audience, medium, and purpose are also of vital importance here.
In addition to choosing a classification method for your maps, you also must decide how many classes to create. It may be tempting to create a large number of classes, as more classes means less simplification of your data, and thus more information conveyed to the map viewer. Unfortunately, the human eye can only differentiate between so many colors. The limit is about a dozen colors for a qualitative map, ten for a diverging scheme, and only eight for a sequential scheme. If anything, these are optimistic estimates—your map reader is likely to be able to differentiate between even less.
View the maps in Figure 4.6.5 below. Looking at the map on the left, can you identify within which class county x belongs? How confident are you that this is the correct answer? What about in the map on the right?
Finally, when classifying your map data, you will have to contend with outliers in your dataset. Consider a county-level map, where one county has double the rate (for example, of people with graduate-level degrees) of any other county in your data. Some classification methods, such as natural breaks or equal intervals, will most likely group this outlier into a class of its own. Other methods, such as quartiles, will simply place it into a group with all the next-highest counties.
There is no rule for which method is best, except that context matters. Is the rate high because that county contains the most prestigious university in the state? In that case, you probably want it to be highlighted on your map. If instead, it is the highest because only five people live there—and two are college professors—you probably don’t. In general, the more data you have, the less likely an outlier is to be noise: this is called the law of large numbers. Whenever possible, however, you should investigate the possible causes of an outlier—there is no substitute for contextual clues.
There are additional ways to classify your data, including by combining methods—for example, using equal intervals for most of the range, and then switching to natural breaks. Methods also exist that consider not just the distribution of data along the number line, but its distribution through geographic space as well. These are beyond the scope and intent of this lesson, but be aware that you may encounter them in the future.
By now, you should feel pretty good about creating a single, choropleth map. Such maps are often requested, designed, and distributed. Yet the power of maps often comes from our ability to compare them. Static maps—all of the maps we’ve discussed thus far—typically only represent one snapshot in time. What if we are interested in how a phenomenon has changed over time, or how it varies between two disparate locations?
View the two maps below in Figure 4.7.1. They are both maps of population density in the United States and are shown at the same scale. At first glance (to non-US residents, perhaps), it might appear that Vermont has a higher level of population density. But take a closer look at the legends -
The legends in the maps in Figure 4.7.1 don’t match—the darkest color, for example, represents a vastly higher level of population in the first map than in the second. How much does population density differ between New Jersey and Vermont?—due to the unmatched legends, it’s really almost impossible to tell.
Using the same data classification scheme for a set of maps is the easiest way to make them directly comparable. For example, the maps in Figure 4.7.2 use the same data, but this time, both legends are equivalent.
This gives us an entirely different view of the data—New Jersey is now visible as obviously more densely populated. Note, however, that this map just took the default classification scheme from New Jersey, and applied it to Vermont, which is still not a good solution. Though it is now easy to compare these states, we are unable to discern which areas of Vermont are more populated than others: they are all simply classified as "less than 562 residents per square mile." Making maps that work well both independently and when compared is a challenging task, and one which we will contend with in Lab 4.
Another important aspect of choropleth—and any—map design is making sure that marginal elements such as legends and labels are well-crafted to support reader comprehension of your map. For example, see Figure 4.7.3. It may seem at first that this legend is too text-heavy: you don’t generally create visual graphics with the intention of asking people to read. However, without this level of detail, the content of the map would be confusing, and many readers would likely misinterpret it.
This map also purposefully places breaks in the data—for example, one break is placed at 24 percent, which is the percentage of all people in the US who are under 18 years old. The break is annotated to inform the reader of this fact—without this annotation, the use of this specific break would not be useful. Additional legend annotations (e.g., “High proportion of AIAN are young”) serve to clarify the map.
Figure 4.7.4 below similarly uses a text explanation to clarify the data mapped. Due to the classification scheme used, the location indicated by the leader line and Prisons* note does not immediately stand out as an outlier. However, given the topic of the map, this explanation is important. We discussed dealing with outliers earlier in the lesson—one option for dealing with a relevant outlier is simply to point it out to your readers via explanatory text. Mapping is all about graphic presentation, but sometimes the best solution is a simple, concise, text explanation.
When using color as a symbol on your maps, your first priority should be to apply it analytically. As stated before: the perceptual structure of your color scheme should match the perceptual structure of your data. You should apply color based on the guidelines previously discussed in this lesson before worrying about choosing aesthetically-pleasing colors, or your audiences’ likely favorite colors, or colors that correspond to the context of the data (e.g., using a green color scheme to create a map about sustainability).
However—when appropriate—adding context to colors in your maps can benefit your readers. See the map in Figure 4.8.1 below. Rather than choosing a traditional sequential color scheme, this cartographer chose to match the colors of the leaves to the colors on the map.
This approach may not always work to best represent the mathematical order of your data classes. But your maps aren’t just about dots along a number line—they represent real-world phenomena. Using color assignments that make sense (e.g., red for negative values), or are customary (e.g., yellow for residential in zoning maps) can improve the clarity and comprehensibility of your maps.
In Lab 4, we will explore different ways of choosing data classification and color schemes for choropleth maps. As a cartographer, you will often have to choose between several of these options - many of which may seem at first glance to be equally appropriate. In Lab 3, we used data from the American Community Survey, provided by the US Census - a commonly-used source of data for statistical maps. In this lab, we use the same data source but focus on a specific variable frequently in focus during public policy debates: health insurance.
The first part of Lab 4 will focus on data classification. There are many ways to classify statistical data on maps, and it is important that you understand them, and be able to defend your choice of classification scheme to others. As we will be not only be classifying data but also adding that data to maps, this lab will also focus on the use of color on maps. Finally, as suggested in the lesson content, we will explore ways of making comparable maps - in this lab, we will be making three pairs of maps.
This lab, which you will submit at the end of Lesson 4, will be reviewed/critiqued by one of your classmates in Lesson 5.
For Lab 4, you will create three pairs of maps, each pair as its own full-page map layout. In total, you will have three separate pages. Two maps will appear on each page. You will also write a short reflection statement about each pair of maps.
A rubric is posted for your review.
More instructions are available in the Lesson 4 Lab Visual Guide.
This is your starting file in ArcGIS Pro. It includes county-level boundary data for the United States. This county-level file has been joined with health insurance data for New England from the American Community Survey (ACS). A state boundaries file is also included – this file is not needed to map the health insurance data, but you may choose to symbolize it to create visible state boundaries on your map.
Within the health insurance data provided in the Lab 4 zipped folder, find two variables you are interested in and their associated universes. For example, if you were interested in uninsured people under 18, your value and universe would be those shown in Figure 4.2 below. (note: this is one variable, you need to choose two).
Paste the four columns you will need "as values" (see Figure 4.3) into the Chosen Data sheet. (Reminder: use something other than just age for your maps). This will eliminate the clutter of the full dataset, giving you space to calculate standardized values from your data. We will use these standardized values to determine class breaks for our first set of maps.
Once you have your two variables of interest (and their universes) in the Chosen Data sheet, use Excel to calculate a standardized column of data for each of your variables. You want to divide each variable of interest by its universe (recall the Data Standardization section in Lesson 4).
Insert a column of 1s and 2s as shown - we will use this to create a dot plot. When you select columns A and B below and insert a scatter plot, this will create a dot plot showing the distribution of your two standardized variables along the number line.
Draw lines with the "insert shape" tool to illustrate where you will be placing breaks in your data. Annotate your lines if you choose the breaks for a reason other than just eyeing the dot distribution. For example, if you place a break at the national average for a variable, annotated this break with a text box explanation such as "US national average." Ex: “national average."
Note that Figure 4.7 is an example of how to draw lines above your dot plot, but these are not good breaks.
We will not be importing our excel data into ArcGIS, as I have already loaded the health insurance data into ArcGIS for you. We only needed the Excel file to decide on what breaks to use for our data classification. Instead of importing standardized values, use ArcGIS to standardize your data for you: make sure the variables you choose match the ones you chose earlier!
You will then manually edit your class breaks to match the ones you drew on your dot plot (use your eye to estimate the values). The screenshot in Figure 4.8 (below) is an example of a screenshot from the Symbology Pane. You will submit a screenshot of the Symbology Pane for both maps in layout one, in addition to an image of your dot plot with annotated breaks.
For these maps, you will be setting a critical class break (e.g., based on the mean of the data) and a diverging color scheme. To create your second pair of maps, choose a diverging color scheme. Then, set a deliberate and useful critical class or break. Once the break is set, you should manipulate the other class breaks manually. As a suggestion, for the other class breaks you could start with the manual breaks you chose for your first two maps, but may need to adjust them to work with this new color scheme. Reference the Lesson 4 reading for ideas and advice on how to choose a critical class or break.
For the third set of maps, abandon your previously-selected class breaks. In this set of maps, you will compare the visual difference between a classed map and an unclassed map. Use the same sequential color scheme for both maps so they can be adequately compared. You should also use consistent line design, etc., so as to not distract from the primary difference of interest - the classification method used. Unlike with the first two sets of maps, you will not be mapping two different variables for comparison here. You will choose just one of the variables from your previous maps, and visualize this variable on both of maps 5 & 6.
For your classed map, choose any of the methods available in ArcGIS Pro – but have a reason why! You will discuss your reasoning for choosing one of these methods in your write-up for this map pair.
Natural Breaks (Jenks): Numerical values of ranked data are examined to account for non-uniform distributions, giving an unequal class width with varying frequency of observations per class.
Quantile: Distributes the observations equally across the class interval, giving unequal class width but the same frequency of observations per class.
Equal Interval: The data range of each class is held constant, giving an equal class width with varying frequency of observations per class.
Defined Interval: Specify an interval size to define equal class widths with varying frequency of observations per class.
Manual Interval: Create class breaks manually or modify one of the present classification methods appropriate for your data.
Geometric Interval: Mathematically defined class widths based on a geometric series, giving an approximately equal class width and consistent frequency of observations per class.
Standard Deviation: For normally distributed data, class widths are defined using standard deviations from the mean of the data array, giving an equal class width and varying frequency of observations per class.
For this lab you will submit three layouts, each containing a pair of maps. You will also submit a write-up document, with a 100+ word explanation of your design (data classification and color) choices for each map pair. Make sure to also design a neat and useful layout - see Lesson/Lab 2 for layout design advice.
Don’t copy this (poor) layout design – use your own knowledge and judgment. Clean up titles, marginal elements, alignments, etc. – use either portrait or landscape, whichever you prefer. Note that elements which refer to both maps (legend; north arrow; scale bar) need only be included once.
Don’t copy this (poor) layout design – use your own knowledge and judgment.
Use convert to graphics to manually improve your legend. Use a text box to annotate your critical class/break!
Don’t copy this (poor) layout design – use your own knowledge and judgment. Remember this map pair uses the same data for each map – it is demonstrating the effects of classification. Your goal should be to make a clean, useful legend for each map - make it look better than the legend design below.
Think about color and what you are mapping. Are you mapping insured or uninsured? Choose colors wisely – what do they represent?
Remember that you can employ text to explain your map! Use text sparingly but effectively – don’t be afraid to use convert to graphics and/or manually edit text and layout elements. When choosing a color scheme as well as when doing your write-up, keep in mind: the perceptual progression of your data should match the perceptual progression of your color scheme.
Credit for all screenshots is to Cary Anderson, Penn State University; Data Source, US Census Bureau.
Congrats on making it to the end of Lesson 4! In this lesson, we learned about color and choropleth maps - two topics that are quite inter-related. During our discussion on color models and human color vision, we talked about how to apply color to maps. We learned how to choose a type of color scheme for a map based on the perceptual progression of our data, as well as how to consider other factors such as map purpose, color accessibility, and data context.
In Lab 4, we made three different pairs of maps. In doing so, we took on the challenge of making maps that work well both independently and when viewed together. We also compared the visual effect of classed vs. unclassed maps, and considered the impact of each method on reader perception of our maps. In building our final map layouts, we utilized knowledge from earlier lessons, such as legend and layout design. As we move forward with the course, the skills we learn will continue to build upon each other. We will design some more interesting map layouts in Lab 5!
You have reached the end of Lesson 4! Double-check the to-do list on the Lesson 4 Overview page [108] to make sure you have completed all of the activities listed there before you begin Lesson 5.
Welcome to Lesson 5! In previous lessons, we discussed and designed several types of thematic maps, including proportional symbol, dot, and choropleth maps. Here, we discuss a more specialized type of thematic map - flow maps. In this lesson, we'll integrate our knowledge of visual variables, map symbolization, and levels of measurement into our discussion of these flow maps: maps that show movement between locations.
Before diving into our flow map discussion, however, we introduce another topic integral to cartography: map projection. We explore the different ways in which we define locations on Earth's surface, the process of creating a map projection, and how our choice of projection alters readers' interpretations of our maps. By the end of this lesson, you should understand the different classes and cases of projections, as well as popular map projections and their characteristics. In Lab 5, we use this knowledge to create custom projections for flow map-based advertisements - a twist intended to emphasize the vast variety of clients and audiences for whom cartographers design thematic maps.
Action |
Assignment |
Directions |
---|---|---|
To Read |
In addition to reading all of the required materials here on the course website, before you begin working through this lesson, please read the following required readings:
If you want to dive into the material a bit further, a good place to start with map projections and learning about their influence on map design, check out this article [109]: Hsu, Mei-Ling. "The role of projections in modern map design." Cartographica: The International Journal for Geographic Information and Geovisualization 18, no. 2 (1972): 151-186. Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allow. |
The required reading is available in the Lesson 5 module. |
To Do |
|
|
If you have questions, please feel free to post them to the Have a question about Lesson 5? Ask here! forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
From a young age, we are generally taught that Earth is a sphere. Images such as those taken from space (e.g., Figure 5.1.1) serve to reinforce this idea. Yet, this is an oversimplification—Earth's actual shape is more complicated than the spherical shape it appears to be.
Due to the centrifugal force created by Earth’s rotation, Earth bulges slightly at the center—it is wider around the equator than from pole to pole. Because of this, a better way to describe Earth’s shape is as an ellipsoid. Ellipsoids which closely resemble spheres are often called spheroids—and as Earth is wider in the East-West direction, the most precise word to describe the approximation of Earth's shape is oblate spheroid (or oblate ellipsoid). In literature about this topic, the terms spheroid and ellipsoid are often used interchangeably. The term you use is less important than your understanding of the general concepts involved.
Though the terms ellipsoid or oblate spheroid better describe Earth’s surface than the term sphere, this is still an over-simplification. Due to gravitational forces and natural landforms, the Earth is not perfectly smooth. A very complex mathematical model has been developed to precisely model Earth's surface, but this is of little practical use in cartography so we will not discuss that model here. The most precise model of the Earth which is of practical use to cartographers is the one called the geoid.
The geoid is defined as “a smooth, undulating surface the Earth would take on if the oceans were allowed to flow freely over the continents—without currents, tides, waves, and so on—which would create a single undisturbed water body” (Slocum et al. 2009, pg. 125). It is an uneven surface which approximates mean sea level across the Earth’s surface by taking gravitational forces into account.
The geoid is constantly changing—due both the ever-changing nature of Earth’s surface (e.g., from continental shifts), and because technological advancements have allowed for more and more precise calculations over the years. The dynamics of Earth and the imprecision of measurement techniques mean that any model of the geoid is only an approximation—and mean sea level itself remains an approximation for Earth’s surface in reality.
Due to the complexity of modeling the geoid and the reasonable similarity between Earth and a regular ellipsoid, we do not use the geoid directly to designate horizontal locations on Earth’s surface. Instead, we use geographic datums—models that describe the locations on an approximation of Earth as a smooth, defined ellipsoid.
Horizontal datums denote locations using a system of longitude and latitude. The network of latitude and longitude lines that appears on a map is called the graticule. Horizontal datums are created using a reference ellipsoid – an ellipsoid whose shape approximates that of Earth’s surface. Not all datums use the same reference ellipsoid. Similar to how maps are designed with a purpose in mind, the specifics of a reference ellipsoid’s shape and position are determined based on the intended use of the datum.
The two illustrations in Figure 5.2.1 below demonstrate how datums differ in their design based on their intended purpose. In Figure 5.2.1 (left), the reference ellipsoid is aligned to closely fit the geoid in one part of the world (Australia). This is a local datum developed for use in Australia, and though the ellipsoid fits other parts of the world poorly, this is acceptable given the datum's intended use. In Figure 5.2.1 (right), the reference ellipsoid more closely fits the geoid overall. This is important for horizontal datums that are used to specify coordinates across the entire globe. The reference ellipsoid in 5.2.1 (right) is also centered at the center of Earth’s mass, which is important for GPS positioning.
Another type of datum is a vertical datum, which is used to specify vertical heights from a base surface approximated using calculations of mean sea level. Vertical datums are important for designing cartometric maps, but we will not discuss them in detail here.
The three most popular datums used in North America are the North American Datum of 1927 (NAD27), the North American Datum of 1983 (NAD83), and the World Geodetic System (WGS84). NAD27 was the first standardized connected system of location points in North America. It was based off the Clarke Ellipsoid of 1866—measurements were made and recorded based on the relative positioning of all locations from Meade’s Ranch in Kansas. NAD83 is the modernized replacement of NAD27 and sought to improve positional accuracy as a result of adding thousands of new benchmarks.
In a time before computers and satellite measurements, why do you think Kansas was chosen to start measurements for the North American Datum of 1927? What role does this location play in GIS today?
NAD83 replaced NAD27 in 1983; its increase in accuracy came from its use of a greater number of benchmarks than NAD27. Still, NAD83, still relied on the human measurement of triangulation from a control points. The World Geodetic System (WGS84) was developed alongside GPS technology, permitted the creation of an accurate worldwide datum. WGS84 is the standard datum used by GPS technologies today, though NAD83 remains popular for non-GPS-based mapping activities in North America. A new horizontal datum for North America, the North American Terrestrial Reference Frame 2022 (NATRF2022), is under development and is planned for release in 2025. More information on this and other new datums is available from the National Geodetic Survey [114].
Historical maps and data often reference the now-outdated NAD27 datum; it is important to be aware of the datum which was used to designate the locations of your spatial data. Datum transformation is the process of re-calculating locations based on a different datum and may be necessary if you are combining datasets that were specified using different datums (e.g., NAD27 vs. NAD83), or if you are hoping to map historical data using a more up-to-date system.
As noted previously, modeling the earth as an ellipsoid or geoid is necessary for Cartometric mapping—mapping that involves the taking of precise measurements. Current GIS software tools (and the computers they run on) are now powerful enough to create projections based on an ellipsoidal Earth without much difficulty. For most thematic mapping purposes, however, conceptualizing Earth as a sphere is close enough.
For the rest of this lesson, we will discuss Earth’s shape as if it were spherical, despite this being an oversimplification. The reason for this is that to create a map—that is, a 2D (flat) rendering of Earth’s surface—we need to represent a 3-Dimensional object on a 2-Dimensional plane. And even with a simple sphere, this is no simple task.
The following lists provides supplemental readings on related topics that you can provide more detail about datums and map projections.
Unlike Earth, maps are flat. Though Earth can also be represented as a globe, globes are inconvenient, expensive, and challenging to design. Maps are much more convenient: they are easier both to produce and to reproduce, and are better suited for displaying detailed data. The process of transforming latitude and longitude values from the 3D earth onto a 2D surface (a map) is called map projection.
In the past, cartographers were tasked with projecting maps by hand. Fortunately, GIS software such as ArcGIS is now able to perform this task of projection for us mathematically. Though manual map projection is uncommon today, terms from this era of map production are still in use and are helpful for conceptualizing how the process of map projection works.
To create a map, cartographers transfer a model of the earth as it appears on a reference globe to a developable surface.
A reference globe is a model of Earth (including landmasses, oceans, and the graticule) at some chosen scale, which is the final scale of the map to be created (Slocum et. al 2009). This projected map is thus modeled from an imaginary scaled-down version of Earth.
A developable surface is a mathematically-definable surface onto which landmasses and the graticule (lines of latitude and longitude) are projected (Slocum et. al 2009). In simpler terms, a developable surface is any surface that can be “unrolled” flat (and thus, create a 2D map). Typically, either a cone, a plane (flat surface), or a cylinder is used. In this next section, we discuss how the choice of a developable surface—among other factors—influences a map projection's characteristics.
Imagine the cone developable surface as a party hat placed on top of Earth. After projection, which locations do you imagine would appear the least distorted on the resulting map? Which would appear the most distorted?
There are many projections to choose from, as well as many options for customizing the projection you choose. Before you decide, it will help to understand the characteristics of different projections. Projections are generally defined by their class, case, and aspect. All three of these characteristics refer to the way in which the developable surface relates to the reference globe.
A projection’s class refers to which developable surface was used to create the projection. Was the developable surface a cone (conic class), plane (planar class/azimuthal), or cylinder (cylindric class)?
Which class of projection you use will depend, among other factors, on the location of the region you intend to map. Planar projections, for example, are often used for polar regions.
As shown by the figure below (Figure 5.4.2), a map will contain no distortion at the location where the reference globe touches the developable surface, and distortion increases with distance from this location.
Even among projections of the same class, there is more than one way to create a projection with the selected developable surface. A projection’s case refers to how this surface was positioned on the reference globe. If the developable surface touches the globe at only one point or line, this is called a tangent projection. If it touches at two, this is called a secant projection.
Figure 5.4.3 illustrates the difference between a tangent and secant projection.
Aspect refers to where the developable surface is placed on the globe. If it is placed over one of the Poles (North or South), this is called a polar aspect projection. If the center is along the equator, this creates an equatorial projection. If the developable surface is placed anywhere else, we call this an oblique projection.
No matter what its class, case, and aspect, all projections have distortion. Just by nature of transforming from a 3D globe to a 2D projection, distortion is inevitable. Different projections, however, have different types of distortion. In the next section, we discuss these differences.
When all else is equal, secant projections have less distortion than tangent projections. Why?
All map projections distort the landmasses (and waterbodies) on Earth’s surface in some way. Even so, projections can be designed to preserve certain types of relationships between features on maps. These include equivalent projections (which preserve areal relationships), conformal projections (angular relationships), azimuthal projections (directional relationships), and equidistant projections (distance relationships). The projection you choose will depend on the characteristics most important to be preserved, given the purpose of your map.
Equivalent projections preserve areal relationships. This means that comparisons between sizes of land-masses (e.g., North America vs. Australia) can be properly made on equal area maps. Unfortunately, when areal relationships are maintained, shapes of landmasses will inevitably be distorted—it is impossible to maintain both.
In Figure 5.5.1 below, shape distortion is most pronounced near the top and bottom of the map. This is because the poles of Earth (North and South) are represented as lines the same length as the equator. Recall that lines of longitude on the globe converge at the poles. When these convergence points are instead mapped as lines, landmasses are stretched East-West, which means that to maintain the same area, landmasses must be compressed in the opposite direction. In the map below, Russia (and other landmasses) are represented at the proper size (compared to other landmasses on the map) but their shapes are significantly distorted.
The projection property of equivalence is perhaps best understood by contrasting its properties with a popular projection that greatly distorts area—the Mercator projection (Figure 5.5.2).
The Mercator projection results in a significant distortion of areas, particularly at locations far from the equator. In order to maintain local angles, parallels (lines of latitude) are placed further and further apart as you depart from the equator. The website thetruesize.com [122] demonstrates this effect.
Despite this, the Mercator is useful for some purposes. It has historically been used for navigation—it is efficient for routing as any straight line drawn on the map represents a route with a constant compass bearing (e.g., due West). This “line of constant compass bearing” is commonly referred to as a rhumb line or loxodrome. The Mercator is a conformal projection.
Conformal projections preserve local angles. Though the scale factor (map scale) changes across the map, from any point on the map, the scale factor changes at the same rate in all directions, therefore maintaining angular relationships. If a surveyor were to determine an angle between two locations on Earth’s surface, it would match the angle shown between those same two locations on a conformal projection.
As mentioned previously, the angle-preserving nature of conformal projections makes them useful for navigation. Any path across Earth that follows a constant compass bearing is called a rhumb line, or loxodrome. Any straight line drawn on a map based on a Mercator projection is a rhumb line. Rhumb lines and loxodromes facilitate navigation, as navigators prefer to follow a straight-line route on the map and set their compass direction accordingly.
Despite their usefulness for navigation, rhumb lines do not show the shortest distance between two points. The shortest point between two points on Earth is called a great circle route. Unlike rhumb lines, such lines appear curved on a conformal projection (Figure 5.5.4). Of course, the literal shortest path from Providence to Rome is actually a straight line: but you'd have to travel beneath Earth's surface to travel it. When we talk about the shortest distance between two points on Earth, we are talking in a practical sense of traveling across or above Earth's surface.
The gnomonic map projection has the interesting property that any straight line drawn on the projection is a great circle route. The gnomonic projection is an example of an azimuthal projection.
Azimuthal projections are planar projections on which correct directions from the center of the map to any other point location are maintained. The stereographic projection is another example of an azimuthal projection. Though only on the gnomonic projection is every straight line a great circle route, a straight line drawn directly from the map’s center is a great circle on any azimuthal projection.
The most common types of azimuthal projections are the gnomonic, stereographic, Lambert azimuthal equal area, and orthographic projections. The primary difference between azimuthal projection types is the location of the point of projection. In Figure 5.5.7 below, a gnomonic projection occurs when the point of projection is Earth’s center. Stereographic maps have a point of projection on the side of Earth opposite the plane’s point of tangency; the point of projection for an orthographic map is at infinity.
Equidistant projections are often useful as they maintain distance relationships. However, they do not maintain distance at all points across the map. Instead, an equidistant projection displays the true distance from one or two points on the map (dependent on the projection) to any other point on the map or along specific lines.
In the azimuthal equidistant projection (Figure 5.5.8, left) distance can be correctly measured from the center of the map (shown by the black dot) to any other point. In two-point Equidistant projection (Figure 5.5.8, right), correct distance can be measured from any two points to any other point on the map (and, thus, to each other). In the example above, those two points are (30⁰S, 30⁰W) and (30⁰N, 30⁰E). These values were supplied as parameters to GIS software while projecting the map—a process called projection customization. When customizing your own projection, you will select locations relevant to your map’s purpose.
Not all equidistant maps are circular in shape. The cylindrical equidistant projection, for example, is equidistant wherein correct distances can be measured along any meridian. When the cylindrical equidistant projection uses the Equator as its standard parallel, the graticule appears to be comprised of grid squares, and it is called the Plate Carrée, a popular map projection due to its simplicity and utility.
Imagine you are planning a flight path and tasked with finding the shortest route from Alaska to New York. Which map would you use? Why? Would the map you use first to draw the route be different than the map you would use while traveling?
So far, we have discussed maps that preserve areal (equivalent), angular (conformal), distance (equidistant), and directional (azimuthal) relationships. As demonstrated by the previous examples, maps that preserve certain properties do so at the expense of others. It is impossible to preserve angular relationships, for example, without significantly distorting feature areas. For this reason, another class of projections exists—compromise projections.
Compromise projections do not entirely preserve any property but instead provide a balance of distortion between the various properties. A popular example is the Robinson Projection, shown in Figure 5.5.10 below. Note on this projection how the landmasses appear more similar in shape and size to what is seen on a globe compared to their appearance on a projection that preserves a specific property entirely (e.g., The Mercator).
Interruption is not a projection property, but interrupted projections can also be useful in some mapping contexts. Interrupted maps, such as the Goode homolosine interrupted projection (Figure 5.5.11), are reminiscent of an “orange-peel” pressed against a flat surface, a common metaphor for map projections.
The interrupted nature of this projection severely distorts (by dividing) water bodies, and so would not be useful for maps related to oceanic data, or those intending to visualize routes across Earth’s (connected) surface. These distortions, however, allow the map to display a more accurate representation of landmasses’ sizes and shapes. Note that while the divisions on the projection shown in Figure 5.5.11 are over water, divisions over land are also possible, though not as popular.
Many projections are available in ArcGIS and other software, some of which are imaginative and fun (e.g., the Berghaus star; Figure 5.5.12) and all of which can be customized to suit a map’s location and purpose. We will talk more about how to select an appropriate map projection in the next section.
There are many factors to keep in mind when choosing a projection for your map. The number of projections available can sometimes seem overwhelming, and as there is no distortion-free map, the selection of any projection involves a trade-off between different properties.
Slocum et al. (2009) provide five suggestions for choosing a projection for a thematic map:
When selecting a projection for your map, your map’s purpose and location should be at the forefront of your decision-making process. Many cartographers have proposed guidelines or tools to assist map-makers in choosing an appropriate projection.
Frederick Pearson (1984), for example, suggested simple guidelines for map projection selection based on the latitude of the area to be mapped. If the map was of an equatorial region, he suggested a cylindric projection. If it was mid-latitude, a conic projection, if it was polar, a planar projection (Pearson 1984). While this is a good starting point, it does not account for the purpose or the map, nor help the map-maker choose between the many projections that exist of each type (e.g., there are many different conic projections).
Some online tools have been developed to help in the projection selection process. One such tool is Projection Wizard, developed Bojan Šavrič (Šavrič, Jenny, and Jenny 2016). It is a web-based tool that suggests projections based on user input of only the intended distortion property (e.g., equal-area), and the location of the map (input via an adjustable map frame).
Projection Wizard is based largely on projection selection guidelines developed by John Snyder (1987), guidelines which are also discussed in detail by Slocum et al. (2009). These sources are listed in the recommended readings for this section—highly suggested if you would like to learn more about this topic.
As noted by Slocum et al. (2009), selecting an appropriate projection requires thinking not only about its objective utility, but about its overall design and what your map’s readers will think of it. Recent research has investigated user responses to map projections. Battersby and Kessler (2012) investigated novice and experienced map-readers’ strategies for comprehending distortion on maps, and found that both groups struggled to correctly specify distortion on maps. Šavrič et al. (2015) focused on user preference and found that many readers tend to favor the Robinson (and similar) projections, and that general map-readers have somewhat different preferences for map projections overall than experienced cartographers.
In addition to attending to projection guidelines and anticipating reader responses, it is often helpful simply to experiment with different projections. The following tools are good resources to explore projection properties and distortion:
Another helpful way to learn is to create a simple map in ArcGIS and practice changing its projection—load simple boundary files (such as those provided for Lab 5) and notice how altering the projection and projection parameters changes the final design.
Two common map projections used in the United States are the Lambert conformal conic and transverse Mercator. The Lambert conformal conic, as its name suggests, is a conformal (preserves local angles) projection that uses a cone as its developable surface. The name “Lambert” is from its inventor—Swiss scientist Johann Heinrich Lambert. Conic projections are particularly useful for mid-latitude regions with primarily East-West extent, such as the United States.
The transverse Mercator projection is a slight alteration of the Mercator projection. Where the Mercator uses the equator as its line of tangency, the transverse Mercator uses a meridian. Figure 5.7.2 below uses the prime meridian as its standard line.
These two projections are used in the State Plane Coordinate System (SPCS), a coordinate system designed for use in the United States. The SPCS is useful for some mapping tasks such as local government planning, as these coordinate systems have been designed to be highly accurate within each zone. Problems can occur, however, when areas of interest cross a zone boundary: this requires that at least one set of data be transformed so that proper GIS analysis can be conducted.
As shown, the transverse Mercator is used in states with a primarily North-South extent (e.g., Vermont, New Jersey) or in locations where the state is usefully divided into multiple North-South extent (e.g., New York). The Lambert conformal conic projection is similarly used for East-West extents. Some states, such as Florida, use both (Lambert conformal conic is used for the Florida panhandle). The oblique Mercator is used only in one case—the Alaska panhandle—as this region has an extent that is neither North-South nor East-West.
Another popular coordinate system is the Universal Transverse Mercator (UTM). The system divides the world into 60 zones, each of which covers six degrees of longitude. The set of zones that covers the US is shown in Figure 5.7.4.
Each UTM zone uses a secant transverse Mercator projection with unique parameters based on the longitudes of its bounds. As the Mercator is a conformal projection, local angles are maintained. Areas and distances are distorted, but the use of secant projections and the somewhat small size of the zones keeps this distortion low – at about 1 part in 1,000. The larger size of these zones means that they are more likely than SPCS zones to cover the entirety of a local area of interest, though recommendations exist for adjusting maps in cases where a mapped area overlaps multiple zones. UTM's worldwide coverage also makes it useful for creating maps that are shared around the world, and it is widely used in military applications.
Choosing an appropriate projection is important for all mapping tasks. Consider, for example, a proportional symbol map. You would not want to use a projection that significantly distorts area—as the intention of such a map is to compare the size of the symbol to the size of its underlying area, this would be misleading.
A map type that we haven’t yet discussed, and to which projection choice can be integral, is a flow map. A flow map is a map that visualizes movement between places—often across large regions, even the entire globe.
Flow maps can be classified into two main types: those that represent origins and destinations, and those that map routes. Origin-destination flow maps show the start and end points (and often the direction) of flows, but do not map out a route. An example is shown in Figure 5.8.1. Flow arrows in Figure 5.8.1 show the direction and magnitude of migration flows, but the route paths are not meaningful. Note, for example, the placement of a large red arrow showing migration from many locations to California. This indicates that many people migrated from these places to California during that time period, but we can imagine that their actual movement covered various routes.
Other flow maps show meaningful routes, such as the flow of traffic, or stream flows. Figure 5.8.2 is an example—instead of focusing on the starts and ends of flows, it maps out a route network. Size is used to visually encode the amount of truck traffic. Though the focus is on truck traffic, traffic volume overall is also visualized in grey, highlighting the difference between routes used primarily by passenger cars and those used for trucking.
Possibly the most famous flow map ever designed was drawn by Charles Minard; it represents the French army’s travel and suffering during the Russian campaign of 1812 (Figure 5.8.3). Edward Tufte, in his influential book The Visual Display of Quantitative Information, described this work as perhaps the best statistical graphic that had ever been created (Tufte 2001).
Another map by Minard (Figure 5.8.5) is more reminiscent of modern flow maps. It illustrates migration flows across the world using multiple visual variables. The achromatic continent fills and boundaries place emphasis on the flowlines as the more important component of the map.
Figure 5.8.5, as well as Figure 5.8.3 (and 5.8.4) above, are examples of aggregating flows to create a more comprehensible map. Figure 5.8.5 shows the magnitude of migration flow between Europe and America, for example, but it does not show the many routes these people likely traveled. Figure 5.8.6 below is an example of the opposite design chose—all paths are mapped. This is appropriate for some mapping purposes, but if there are many routes, this makes the map more challenging to read.
Figure 5.8.6 also differs from the other flow maps shown above in that it does not visualize any data except the flight paths and endpoints. When creating flow maps, whether you map entire routes or just origin-destinations, and whether you chose to visually encode additional data, such as with size or color hue, will depend on the intended purpose of your map.
Flow maps can also be combined with other types of thematic maps, such as proportional symbol or choropleth maps, to show multiple sets of data. Figure 5.8.7, for example, combines a qualitative choropleth map with directional flows.
In Figure 5.8.7 above, what visual variables are used? What levels of measurement are used to map the flows?
Critique #3 will be your second critique involving a peer review. For this critique, you will be reviewing a colleague's map from Lab 4: Color and Choropleth Mapping in Series. During that lab, you put significant thought and effort into classifying your data and applying color - now you will appraise another's work instead of your own. This new perspective is likely to be beneficial to you both while you are writing the critique, and later, when you review the feedback provided to you by one of your peers.
Your assignment includes writing up a 300+ word critique of an assigned classmate’s Lesson 4 Lab (as assigned).
In your written critique please describe:
As Lab 4 included three map-pair map layouts as the deliverable, the best way to approach this critique is to write about one well-done element and one suggestion for improvement for each pair of maps. You may stray slightly from this format if you are particularly interested in one of the map pairs (layouts), but, please, at least briefly mention all map pairs in your critique.
Your critique should be as much about reflecting upon things well-done as it is about suggesting improvements to be made. In your discussion, you should connect your ideas back to concepts in Lesson Four; you may also reference concepts from earlier lessons where relevant.
Please list the student name of the map you have been assigned at the top of the page.
A rubric is posted for your review.
You will work on Critique #3 during Lesson 5 and submit it at the end of Lesson 5.
Step 1: Upon notification of the Peer Review (Critique), go to Lesson 4: Lab 4 Assignment. You will see your assignment to peer review. (Note: You will be notified that you have a peer review in the Recent Activity Stream and the To-Do list. Once peer reviews are assigned, you will also be notified via email.)
Step 2: Download/view your classmate's Lab.
Step 3: Write up your critique using the prompts above in a Word document. Be sure to also review the rubric in which you will be graded for Critique #3 for more guidance. Save your Word document as a PDF. Use the naming convention outlined below.
Step 4: In order to complete the Peer Review/Critique, you must
- Add the PDF as an attachement in the comment sidebar in the assignment.
- Include a comment such as "here is my critique" in the comment area.
- PLEASE DO NOT complete the lesson rubric as your review, award points, or grade the map you are critiquing. Even though Canvas asks you to complete the rubric, PLEASE DO NOT COMPLETE THE RUBRIC OR ASSIGN POINTS/GRADE.
Step 5: When you're finished, click the Save Comment button. You may need to refresh your browser to see that you've completed the required steps for the peer review.
Note: Again, you will not submit anything for a letter grade or provide comments in the lesson rubric.
.
In Lesson 5, we discussed map projections and projection characteristics. We also discussed how to choose a map projection based on your map's intended location, scale, and purpose. It can be challenging, however, to really understand how the choice of a projection alters your map without trying it out for yourself. In Lab 5, we will be creating three map layouts that visualize flight data as flowlines. This ties together both of the topics in Lesson 5 (flow mapping, projections), and provides a practical demonstration of the influence of map projections in small-scale thematic mapping.
For good measure, we will design each of these map layouts as advertisements: encouraging creative design and adding emphasis to the importance of map purpose and audience in choosing projections for maps. Recall from this week's required reading, Mark Monmonier's discussion of Maps that Advertise. Your challenge this week is to create map layouts that are both scientifically-appropriate and engaging to your intended customers - the readers of your maps.
For Lab 5, you will use the provided data to create three different map layouts, each of which is an advertisement for LHR airport.
A rubric is posted for your review.
Further instructions are available in the Lesson 5 Lab Visual Guide.
Throughout this lab, keep the following statement in mind:
"The projection you choose will depend on the characteristics most important to be preserved, given the purpose of your map."
This is your starting file in ArcGIS Pro: It contains boundary, flight, and tourism data. The flight data is in table form - we will be using these data tables to create flight paths and visualize them on the map.
The primary flight data table is the one shown below - it contains a full day of flight data (Oct 22nd, 2018). Listed in the table are all locations which had a flight arrive from, or depart to, London Heathrow Airport (LHR). We will not differentiate between arrival and departure flights in this lab.
The count of flights to or from this location is listed in the Count_Num field. For the purposes of this lab, we will assume that October 22nd is representative of an average day at LHR, and thus use this dataset as a proxy for LHR’s “daily” flight data. You do not need to mention October 22nd anywhere on your maps.
Flight data source [142].
In our flight data, we have lots of origin-destination data. We want to visualize these data as flows on our map. For this, we use the XY to line tool [143]. Think carefully about the fields you choose for each parameter when running this tool. If you do it incorrectly the first time, don't worry - re-think and re-do.
For each map layout in this lab, you will be creating a customized map projection (use the Project tool). Reference the projection lesson and consider each advertisement's goal/purpose to help you decide which projection to choose/customize for each map. You may want to try adding the map to a layout at this stage of the lab to decide if you like it. Remember that you will be asked to defend this choice in the reflection you submit with this lab.
You may need to try a few different projections or customization parameters to find a map projection you are happy with. When you have settled on a projection, use the project tool to project your flowlines to match the map's projection. As shown below, ArcGIS makes this pretty easy.
Recall that we will be visualizing flight paths based on their length. We can use the Shape_Length field which ArcGIS automatically calculated for us from our origin-destination data to do this. Note that before projecting these lines, the Shape_Length field will not contain meaningful values.
Once your flight paths are projected, the Shape_Length field will be calculated in meters.
As you may have noticed, the flight path data doesn't contain any location names or flight count numbers. We'll need to join the original flight data table to the flight path data to get all our data in one place.
Use external research or the data distribution to decide on classifications for short vs. long flights, etc. (3+ classes). Recall considerations for data classification from Lesson/Lab 4.
You may use any or multiple visual variables of your choice to symbolize your flowline data - size, value, etc... as long as it is appropriate given the perceptual structure of your data, you can be creative with it.
Add your map to a layout: create a catchy title/subtitle and customize your legend. Add a graticule (ArcGIS Pro calls this a “grid”) if you wish.
Make sure all layout elements are neat and orderly – “convert to graphics” will likely be helpful. Keep in mind lessons from previous labs: legends and any explanatory text should be clear, etc.
You’ll want to create three new maps (four total) to separate the flight types (morning, afternoon, evening, night). If you prefer, you can do a Save-As and keep work done on this project separate form the previous one. In any case, save frequently!
You can drag the tables onto their appropriate map from the Contents Pane.
Use creativity, appropriate visual variables, and good design in this ad as well! Remember the goal of this layout - highlighting that Heathrow can fit anyone’s schedule.
For the third advertisement, we will add additional data to our map to demonstrate to the reader that flights from LHR go to desirable locations. Choose a field such as “International Tourist Arrivals 2017” that makes sense to use in an ad about air travel. Visualize this data on your map how you choose - remember that you will still be visualizing the flight paths. You may use the same flight path layer from Map #1, but you will need to re-project it to match your projection for Map #3.
Choose and customize a projection you haven’t used yet – be prepared to write about the reason for this selection. Think about your map type and purpose.
Ensure that both your flight path and other thematic data is included in your layout - below is just an example of how you might symbolize this data, but there are many other possible ways. If you do not want to use tourism data, you can use a field that was automatically imported with the Natural Earth boundary data such as GDP. Consider how you will visualize null values.
Credit for all screenshots is to Cary Anderson, Penn State University; Data Source: Flightradar24, Natural Earth.
Welcome to the end of Lesson 5! In this lesson, we discussed the complex process of modeling Earth's surface, and how concepts such as reference ellipsoids and datums relate to the map projections used by cartographers every day. During our discussion of characteristics of map projections, we focused on the appropriateness of various map projections for different mapping tasks: based on a map's location, scale, and purpose. Finally, we connected these ideas to a new thematic mapping technique - flow mapping. Though projection choice is often particularly consequential in flow map design—due to the nature of the data visualized, and to the large regions such maps often depict—it is an important consideration in many mapping projects. You will often have to select an appropriate map projection when making other kinds of thematic maps, including proportional symbol, dot density, and choropleth maps.
In Lab 5, we explored the effect of projection selection on small-scale thematic map design while creating map-based advertisements for London Heathrow Airport (LHR). We designed these maps using prior knowledge of visual variables and symbols on maps, and put together neat, useful layouts intended to appeal to our map readers. Prepare for another creative real-world mapping task in Lab 6!
You have reached the end of Lesson 5! Double-check the to-do list on the Lesson 5 Overview page [144] to make sure you have completed all of the activities listed there before you begin Lesson 6.
Welcome to Lesson 6! Last lesson, we talked in-depth about map projection: the process of transforming Earth's three-dimensional surface into a form that can be depicted on a flat map. Earth's terrain poses a similar challenge - how can we depict the intricacies of Earth's surface on a two-dimensional piece of paper or computer screen? Fortunately, just as with the challenge of projections, cartographers have been designing creative solutions to this problem for many years. In this lesson, we'll learn about many techniques that exist for modeling Earth's terrain. These include both oblique and vertical map views, contour maps, and physical models. We'll also talk a bit about how different terrain layers are built-in GIS software, and the importance of balancing the visualization of terrain with other map data, such as political boundaries, roads, water features, and trails.
In Lab 6, we'll put all this together to create a trail run map for an imagined event, The Paradise Valley Trail Run. You'll generate and design terrain layers, overlay additional base and thematic data, and use your knowledge of symbol and layout design to create a map that would be helpful to runners and their supporters. Let's get started!
Action | Assignment | Directions |
---|---|---|
To Read |
In addition to reading all of the required materials here on the course website, before you begin working through this lesson, please read the following required readings:
Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allow. |
This week's reading is provided in ebook form through the Penn State library system. |
To Do |
|
|
If you have questions, please feel free to post them to the Lesson 6 Discussion Forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
In Lesson Five, we discussed map projections—the act of transferring the 3D earth onto a 2D map. In this lesson, we discuss a similar problem—representing Earth’s three-dimensional terrain surface on a 2D map.
When artists depict three-dimensional landscapes, they commonly use an oblique view. See the example painting in Figure 6.1.1—the perspective of the drawing makes the landscape appear three-dimensional, though it is only a two-dimensional piece of art.
Whether in an artists’ rendering (Figure 6.1.1), photograph (Figure 6.1.2), or digital model, the oblique perspective is effective in its realism: it depicts what might be seen by a person on or near the ground.
Though the oblique view creates a favorable artistic impression, it has its disadvantages. First, this perspective inherently obscures some of the landscape—mountains and similar heightened features hide the land behind them. Secondly, oblique views are often constructed by exaggerating the height of landforms, so as to create an interesting visual depiction. This can make between-map comparisons challenging, and cause issues for cartographers hoping to take accurate measurements with such maps.
To account for these shortcomings, several vertical view techniques for depicting terrain were developed. Figure 6.1.5 shows a topographic map from the United States Geological Survey (USGS), which depicts a section of Acadia National Park. Topographic maps are maps that quantitatively depict terrain, typically with contour lines. Contour lines on a map connect points of equal elevation, and when drawn, they visualize hills, valleys, and other landforms. In the next sections, we discuss in further detail techniques for using both oblique and vertical map views to represent Earth's terrain.
Visualizing three-dimensional terrain without obstructing parts of the landscape has been a challenge in cartography for centuries. Can you think of a modern mapping technique that presents similar problems and challenges for map-makers and readers?
Despite the challenges involved with accurately depicting and visualizing all of the landscape with an oblique view, such views are still useful in some contexts. For some map uses, for example, a detailed view of a small part of the terrain may be more useful than a view from above of a wider area. As with all maps, attention to audience, purpose, and medium is important, and cartographers take these factors into account when deciding how to best represent terrain on a map.
One technique, used commonly in Geology to show underground rock or soil properties, is the block diagram.
Block diagrams show the surface of the landscape as well as underground structures and materials. This gives them a natural advantage over vertical-view maps if the goal of the map is also to visualize both above and below Earth’s surface. The disadvantage of block diagrams is that they cannot depict all sides of the terrain. In Figure 6.2.1, for example, it is unclear whether the composition of underground materials behind the diagram matches that shown in the front. These diagrams are also more challenging to create than traditional maps, though new software developments continue to make this process easier.
Imagine viewing a block diagram such as the one in Figure 6.2.1 in an interactive web environment, rather than on paper. How might this alleviate some of the problems caused by the oblique view?
Panoramas are wide-angle views of an area and another popular technique for visualizing terrain. Several maps we saw in the first section, such as Figure 6.1.3, are panoramic maps. The map in Figure 6.2.2 is available from the Library of Congress—if you are interested in these types of historical maps, the LoC is an excellent source to explore.
The birds-eye perspective often given by panoramic maps provides an easily-comprehensible view of the landscape to the map user. Hills and valleys, for example, appear as they would to an observer in the real world, and thus their recognition requires no prior knowledge of cartography. Despite this, these maps are uncommonly used for scientific purposes as they do not show a geometrically-accurate view of the landscape.
The map in Figure 6.2.3, for example, is a beautiful depiction of the mountains in Wrangell-St. Elias National Park. But if a map reader were to take measurements from this map, they would not be correct. Not only does the oblique view complicate measurement tasks with such maps, but mountain heights are typically exaggerated—not drawn to scale.
Draped images are a form of oblique view maps that have recently become more popular due to the increased availability of satellite imagery and advances in 3D visualization software. They are created by—in essence—draping a remotely-sensed image over a 3D digital terrain model. An example is shown in Figure 6.2.4.
The combination of remotely-sensed data and terrain visualization in draped images can be particularly useful for analyzing a combination of terrain and surface characteristics (e.g., for research on forest fires or ecological suitability).
The oblique view, when compared to the vertical view, provides a more intuitive view of Earth’s landscapes. However, there is an even more intuitive way to model landscapes—with physical 3D models.
Physical models have been around since the time of the Ancient Greeks, but the time and expense required to create such models has sharply decreased in recent years due to the advent of new computer modeling techniques and 3D printing capabilities (Slocum et al. 2009). This has led, as you might imagine, to a recent increase in the popularity of such maps.
Physical representation can be combined with other terrain visualization techniques. The USGS, for example, produces topographic raised relief maps, such as the one in Figure 6.3.2. These maps combine the contour mapping technique with a haptic representation of terrain—creating engaging as well as useful maps.
Another new technology, augmented reality (AR), has become popular for creating realistic and dynamic physical models of landscapes. Shown in Figure 6.3.3 below is an augmented reality sandbox, which draws contour lines and hypsometric tints by detecting the shape of the landscape as molded by sandbox-users.
A similar sandbox is available at UCLA. Watch this video, UCLA's Augmented Reality Sandbox [156], for an exciting demonstration of this technology.
We will talk more about applications of augmented reality and similar technologies (e.g., virtual reality, mixed reality) later in the course.
Techniques for depicting terrain from directly above were developed for two primary reasons. First, the oblique view inherently hides some map features; a vertical view, by contrast, offers a view of all landscape features within the map frame. The vertical view also allows the map maker to position features appropriately in geographic space—thus providing concrete spatial information, rather than a more artistic visual representation (Slocum et al. 2009).
In the vertical view, terrain is typically represented with contour lines. Contour lines drawn on a map connect points of equivalent elevation. Figure 6.4.1 demonstrates how contour lines relate to the landscape from which they are derived—note that the bottom image is a 2D rendering of what is presumed to be a regularly-shaped mountain feature.
As demonstrated by Figure 6.4.1, gentle slopes are represented on contour maps by lines spaced farther apart than steep slopes. This is because elevation values change more quickly across steeper slopes, meaning that contour lines will need to be drawn more often (across the same map distance) to accurately represent the terrain. Figure 6.4.2 below shows a topographic map with markings to denote gentle and steep slopes, as well as valleys, hills, and ridges.
A map’s contour interval is the change in elevation (typically in meters) between drawn contour lines. This is a form of sampling (e.g., every 20m), meaning that topographic maps do not display every possible contour line, but rather display (as all maps do) a simplified view of the landscape.
In addition to mapping elevated features such as hills and mountains, contour maps are also useful for depicting underwater terrain. While topographic maps visualize elevations above sea level, bathymetric maps depict elevations below sea level.
On topographic maps, increasing values indicate higher elevations. Bathymetric values—as they also represent a distance from sea level—increase in the opposite direction. So just as the highest values on topographic maps represent the highest mountains, the highest bathymetric measurements represent the deepest depths of the earth’s oceans.
Despite their usefulness in accurately depicting terrain, contour lines do require some prior knowledge for their proper interpretation, as they do not present an immediately intuitive view of the landscape. To mediate this, cartographers have developed innovative methods for artistically depicting terrain on vertical-view maps using additional elements of design.
One popular method is Tanaka’s method (Tanaka 1950), often called Tanaka contours. Tanaka contours assume that the map is being illuminated by a light source from some direction, and contour lines are drawn lighter (i.e., illuminated) and thinner when facing the light source, and darker (i.e., in shadow) and thicker when perpendicular to the light source. The result is a contour map wherein the form of the landscape is more intuitively depicted (Figure 6.4.5). Ridges and valleys are far less likely here to be confused.
A similar but simplified method called illuminated contours was developed by J. Ronald Eyton (1984).
This method, shown in Figure 6.4.6, varies lightness as in Tanaka’s technique but does not vary line thickness. Contrary to Tanaka’s approach, which was applied manually, Eyton (1984) developed his method in the early days of computerized mapping—he used consistent line thickness to reduce computation time.
Other techniques for designing contour maps have been developed by other cartographers. You are encouraged to explore the recommended readings or search the web on your own to learn more about these techniques.
A mostly-outdated alternative to contour lines called hachures also exists. Hachures are created by drawing a series of lines perpendicularly to contours. The spacing between hachures are drawn proportional to the slope—steeper areas are highlighted by increased density of these lines (Slocum et al. 2009). A hachure-like technique can also be used to manually create shaded relief (a visually-appealing and artistic depiction of landforms), but its traditional purpose was to show a geometrically-correct depiction of slope.
Shaded relief is commonly added to maps to give the reader a more intuitive impression of landform shapes. It presumes the existence of an imaginary light source and displays shadows over landforms accordingly, giving the illusion of depth. An example is shown in Figure 6.4.8.
The light source imagined in shaded relief mapping comes traditionally from the upper-left of the map (Northwest, assuming a North-up map view). At first, this might seem inappropriate—the sun rarely shines onto the Earth from a Northwestern direction, at least in the locations where most people live. This convention does not come from the earth sciences, however, but instead from guidelines in art developed in response to the realities of everyday life at the human scale.
Humans are used to illumination from the sun—as well as other light sources (e.g., lamps, overhead lighting)—coming from above our heads. As most people are right-handed, an upper-left light source ideal for writing. Even left-handed people typically write from left-to-right and top-to-bottom, due to the left-right convention of most languages. Figure 6.4.9 demonstrates the appropriateness of this upper-left light source.
We have become so accustomed to this location of light that light projected from other directions (e.g.. from underneath) results in features looking incorrect to the human eye. Imagine someone holding a flashlight underneath their chin in the dark—the reason their facial features appear so strange is that we are accustomed to seeing them lit from above.
Figure 6.4.10 below shows how changing the azimuth direction of an imagined light source can create confusion in the interpretation of landscape features. Both below maps depict the same location, and a valley exists within the yellow box on each. Left, the valley is shown via traditional Northwest illumination. When the map is illuminated from the Southeast (right) the valley now appears inverted—it looks like a ridge.
Much of cartography is about understanding not only the analytical elements of landscapes and map design variables, but human perception. The Northwest oblique light source convention is an excellent example of how cartographers have developed their techniques with this understanding in mind.
Before the widespread use of computers and GIS for map-making, terrain visualization techniques such as hachures were drawn by hand, and elevation values were approximated using photographs and survey data. In modern cartography, almost all terrain layers begin with one map layer—a digital elevation model (DEM). Though you likely often see DEMs with additional design elements such as color tints and shaded relief, DEM data is actually as simple as shown in the image in Figure 6.5.1 below.
DEMs are raster, or grid-based data. Each grid cell (also called a pixel) in a DEM image has a single value, which corresponds to its elevation. In Figure 6.5.1 for example, the values closest to white are the locations of highest elevation at this location. Using GIS software, DEM data can be used to easily create additional terrain layers—the most common being hillshade, curvature, and contours.
Hillshade is a term often used interchangeably with the term shaded relief discussed earlier. Hillshade is a greyscale raster data layer which uses lightness to imitate the highlights and shadows that would be cast by a hypothetical oblique light source. The highest values in a hillshade layer, then, are those which would be met with the highest levels of illumination from the light source.
Contour lines, as discussed in the vertical views section lesson, connect points of equal elevation across a terrain surface. The density of lines across the map depends on the slope of the terrain—steeper slopes result in lines being drawn closer together. When creating a contour map, you choose what contour interval to use on your map. Theoretically, an infinite number of contour lines can be drawn on any map. Cartographers typically consider multiple factors when choosing a contour interval, including the scale of their map and the steepness of the terrain.
A common technique when symbolizing contour lines on maps is to draw index contours—contour lines that are more visually prominent—at less frequent intervals. Often, to avoid map clutter, only these contour lines are labeled. Map readers can then use the lines between them, called intermediate contours, to interpolate elevation values between them.
Digital Elevation Models can also be used to generate curvature layers, such as the one shown in Figure 6.5.5. Curvature is often referred to as “the slope of the slope.” In mathematical terms, it represents the second derivative of a terrain surface (Muehrcke, Muehrcke, and Kimerling 2001). We will not go into the technical details of how this layer is calculated—the important thing to know for map design is that curvature is excellent for showing inflection points in a surface—sharp ridges and deep valleys. In this way, adding a curvature layer can add additional visual interest to your terrain map.
Viewed individually, none of these layers are very interesting. However, with just a digital elevation model from a source such as The National Map, you can generate several different terrain layers and use layer transparency, color, and other design elements to create imaginative depictions of Earth's terrain. Though terrain visualizations are typically used as a base layer for thematic or general-purpose map data, making maps just of Earth's terrain and experimenting with new, creative designs can be quite fun.
Though terrain layers can be used to make fun and interesting map designs, terrain is rarely the sole element on a map. USGS topographic maps, for example, depict much more than just contour lines across the landscape—they also include political boundaries, streets, water features, and more. This is particularly challenging in urban areas, as demonstrated by the map in Figure 6.6.1, located in Manhattan, NY.
Even when terrain is the main feature of interest, such as in the thematic map in Figure 6.6 2 below, design adjustments must be made to ensure the terrain is visualized appropriately given the map’s projection, level of detail, other visual variables (here, color), and background.
Some types of maps more frequently contain depictions of terrain than others. As designing a good terrain base layer typically involves significant effort—and makes map symbol design more complicated—terrain is typically left off of maps when it is considered irrelevant, such as in thematic maps of political or social data. In some maps however, (e.g., maps of ski trails), terrain visualization is essential. Most maps fall somewhere in between.
Whether or not you decide to depict your location’s terrain—and how detailed that design will be—will depend, as with most design decisions, on your map’s intended audience, medium, and purpose. You will likely also need to take other constraints into consideration (e.g., availability of data and time).
Google maps (maps.google.com) offers users the option of replacing the default Google basemap with a map that visualizes terrain. What use cases can you imagine for routing over such a basemap, rather than the simpler standard map?
So far in this course, we have been working primarily with vector data. Though scale is an important consideration in all mapping tasks, working with raster data such as Digital Elevation Models presents a unique set of challenges for data management and design.
When mapping terrain, it is important to use elevation data that is appropriate for the scale of your map. The image in Figure 6.7.1, for example, appears pixelated and blurry. The resolution of the data used (1-arc-second) is too coarse for creating a clear image at this scale.
The solution to this is, as you might have guessed, to use higher-resolution data. See, for example, the map in Figure 6.7.2. The scale of this map is the same as in Figure 6.7.1, but the finer-grained data results in a much clearer image.
It is important to note that the answer is not to always use the highest-resolution data you can find. The map in Figure 6.7.3, for example, shows a 1-arc-second DEM: the same as used in the blurry image in Figure 6.7.1. At this new scale (1:120,000) this coarser data is quite appropriate. To understand the difference in scale between these maps, note that the extent of the maps above (6.7.1 and 6.7.2) is shown by the blue extent indicator in Figures 6.7.3 and 6.7.4 below.
Raster data is much more space-intensive than vector data, and high-resolution raster data means particularly large file sizes. Using coarse data when appropriate will keep you from unnecessarily filling up all the space on your computer. This is not the only reason for not always using high-resolution DEM data, however. Using data that is too fine for a particular scale can result in undesirable visual effects, similarly to how using data that is too coarse can lead to a very pixelated image. Figure 6.7.4 is an example of a map created with terrain data that is a bit too unecessarily-detailed for its scale.
The good news in this second example is that DEMs can be simplified: GIS software can be used to re-sample and generalize terrain data. As with all data processing tasks, however, it is not possible to go in the opposite direction. The only way to create a more detailed terrain map is to collect more detailed data.
In this lab, you will be creating a map of the (imaginary) Paradise Valley Trail Run in southern San Francisco, California. Imagine the final map will be handed out in race packets - what do trail runners and their supporters want to see? As the race takes place over hilly terrain, you will first design the terrain backdrop of the map, and then add overlay data such as route paths, water stops, and general base data. Finally, you'll put it all together in a layout with an elevation profile for the 10K route and map marginalia.
This lab, which you will submit at the end of Lesson 6, will be reviewed/critiqued by one of your classmates in Lesson 7 (critique #4).
For Lab 6, you will be creating only one map layout, though it will contain several different elements: the primary map, an inset map, an elevation profile, and marginal elements (scale bars, north arrows, text, and legend).
A rubric is posted for your review.
Note: While Paradise Valley is a real place in California, data related to the Paradise Valley Trail Run in this lab was invented and built by the course author. Any existence of a real event with this name or in this location is coincidental. The Resources menu links to important supporting materials, while the Lessons menu links to the course lessons that provide the primary instructional materials for the course.
Please refer to Lesson 6 Lab Visual Guide.
This is your starting file in ArcGIS Pro. It contains data for the Paradise Valley Trail Run, as well as base data (e.g., boundaries, transportation) and a Digital Elevation model (DEM). Your goal is to turn this data into a map for trail race participants and their supporters.
Your first goal in this lab is to use the included DEM to generate additional terrain layers. Create three terrain layers: Hillshade, Contours, and Curvature.
The default settings/parameters provided by ArcGIS are ok for generating the Hillshade and Curvature layers. Make sure your output is saved to the geodatabase for the current project (Lab6_data.gdb).
You will need to choose an appropriate interval for your contours - if you don't like the result, you can always choose a new interval and run the tool again.
Keep your terrain layers organized in the "terrain" layer group in the contents pane - think about your layer ordering, and don't be afraid to re-order layers as you go! Use the transparancy slider so you can see multiple layers at once - all of your terrain layers should contribute to your design.
Try out different symbology methods and color schemes. A simple stretch sequential color scheme (often greyscale) tends to work best for hillshade and curvature, but you can be a bit more creative with the DEM. Right click on a color scheme to reverse it if needed. Remember that higher hillshade values represent greater illumination - so unlike with most map data, higher values should be paired with lighter color. Keep your design subtle enough for your thematic (race info) data to show up on top. This map design is all about balance.
Symbolize the transport, hydro, and boundary layers as appropriate for this map’s purpose. Reference previous labs (particularly 1 and 2) for basemap design ideas. Remember you can create new label classes using SQL! This base data should be visible over the terrain data, but not be so overwhelming so as to detract from the data about the Paradise Valley Trail Run.
Choose line width, color, etc. to symbolize the two race routes. Think about how you can you display these two (overlapping!) routes at once. Design labels for water stations, route markers, and Start/End points. The Gallery may have helpful ideas for your point symbol designs, and there are many ways you can customize them yourself. Explore the available options. You may also want to look at running or trail maps on the web for ideas - but note that some that you find may not be well designed!
Once you are happy with your primary race map, you're ready to start experimenting with layout designs and adjusting your map scales. To design your inset/locator map, it is recommended that you follow the familiar "Save-As map file" and re-import procedure illustrated below. Save a copy of your map, then import it into your map project. You can then alter the design so it works as an inset map.
The Navigator can be used to change a map’s orientation when the map is activated. Remember that your primary map cannot be directly North-Up for this project!
We want to create an elvation profile to help trail runners anticipate the difficulty of the race. To do this, we will be using ArcGIS Pro’s Interpolate Shape [174]tool. This tool turns a 2D line feature into a 3D line feature based an input DEM or other surfaces. We will use this 3D line feature to create an elevation profile. You do not need to create an elevation profile for the 5K route, but you may do so if you choose.
Once you have created a 3D line, you can use this line to create a profile graph. As noted below, the design of your profile graph can be edited. You can also wait and edit the design as you work on your map layout.
Your profile graph will cover a slightly different horizontal distance than in the screenshot below - this is ok!
An important part of route maps like this is to inform the reader of their direction of travel! There are many options for adding directional arrows to your map - two are listed below. You may design your arrows any way you want as long as you do not use any software other than ArcGIS Pro.
Option #1: Use the Edit [175] tab to create arrow features by drawing new lines. An empty “Arrows” feature class has been added to the map for you to facilitate this method. Use the editing toolbar to finish or discard map feature changes in this layer. And always save your edits!
Option #2: Manually add arrows to your map via the map’s layout shape/line tools.
ArcGIS has tools for adding arrows and editing graphics, but is not fully-fledged graphic software (e.g., Adobe Illustrator). Keep this in mind as you decide which of options #1 and #2 for adding arrows works best for you. You might also try them both out and see which works best for your map.
Insert your 10K elevation profile into your layout. (But note that you can keep the old 2D route for your map design).
Map routes, stops, and marker locations are approximate. You may alter them slightly if you would like. Reference the lesson and previous labs for ideas. Check the lab assignment for a list of specific requirements and ask questions in the discussion forum. Don't forget to add an extent indicator and marginal elements (e.g., scale bars, north arrows). Feel free to customize your layout and map elements creatively!
Credit for all screenshots is to Cary Anderson, Penn State University; Data Source: The National Map.
You've reached the end of Lesson 6! This lesson, we discussed the many techniques available for visualizing Earth's terrain, including vertical views (e.g., contour lines, hachures), oblique views (e.g., panoramas, draped images), and 3D physical models. We also explored the terrain layers available to be generated and designed in ArcGIS and similar software, and talked about the importance of DEM resolution (scale) for terrain-mapping projects.
In Lab 6, we put all this together with concepts from earlier lessons. We built a map for an imagined trail run in San Francisco, which involved the design of base, thematic, and underlying terrain data, as well as the composition of a neat, useful, and visually-appealing layout. This kind of mapping task is quite common—cartographers must often combine techniques from many different aspects of map design in their work.
Another important aspect of this lab was our focus on the intended map-reader: someone running a trail race, or cheering on a participating friend or family member. We'll talk more in-depth about map readers (and map users, in the case of interactive maps) in upcoming lessons. How can we design maps so that they best communicate our data, or assist their readers in making better decisions? Continue to Lesson 7 to find out.
You have reached the end of Lesson 6! Double-check the to-do list on the Lesson 6 Overview page [176] to make sure you have completed all of the activities listed there before you begin Lesson 7.
Welcome to Lesson 7! During the course so far, we have discussed many ways in which cartographers symbolize data on maps. We used visual variables to create category and order for basemap and label design, compared proportional symbol, dot, and choropleth maps, and visualized flowlines and terrain. In most cases, our maps have focused on one data variable (e.g., % of people with health insurance), or layered different kinds of data (e.g., race routes layered over terrain). This week, we introduce multivariate mapping—maps that visualize more than one data attribute at once.
Following our discussion of multivariate maps, we introduce a special type of data—uncertainty. When mapping predicted flood zones, for example, we might want the reader to understand not only the predicted flood values across the map, but their associated uncertainty—how certain those values are to reflect reality across different locations. As uncertainty plays a pivotal role in decision-making, we close out our discussion of uncertainty visualization with a short summary of its influence on decision-making with maps. In Lab 7, we explore both multivariate data and uncertainty visualization techniques while creating maps for a new imagined set of decision-makers—a group of policy-makers at the United Nations—using data from the World Happiness Report.
Action |
Assignment | Directions |
---|---|---|
To Read |
In addition to reading all of the required materials here on the course website, before you begin working through this lesson, please read the following required readings:
Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allow. |
The required reading material is available in the Lesson 7 module. |
To Do |
|
|
If you have questions, please feel free to post them to the Lesson 7 Discussion Forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
So far in this course, we have discussed many different ways of symbolizing data using visual variables. Our focus has been primarily on univariate maps—maps that show only one thematic data attribute.
This is a good start, but cartographers often wish to map more than one variable in a thematic map. This is called multivariate mapping. When creating multivariate maps, you should think about the best way to map each individual variable, as well as how you can best combine them to suit your maps audience, medium, and purpose.
The map in Figure 7.1.1 is a multivariate map. The map visualizes two variables at each location—rent prices, and the number of Section 8 vouchers. These variables are each mapped appropriately individually: first, rental prices are visually encoded with a sequential color scheme, a good symbolization choice for normalized data such as rates. The number of Section 8 vouchers at each location is mapped with size, an appropriate visual variable for mapping count data. Together, these symbols work together to visualize this housing data from Portland.
Note that the legend in Figure 7.1.1 is more complicated than many of the legends that we’ve seen so far. The format shown—one variable along the x-axis, and one along the y-axis, is common in bivariate maps, or maps that display two data variables. It not only explains how to data is encoded, but helps the map reader to understand how the data are related to each other. The more complicated a map becomes, the more challenging it will be to design a useful legend. Legend design is an important task however, as your legend is key for proper reader interpretation of your map. The map in Figure 7.1.2 below uses short text blurbs to assist the reader in this interpretation.
As we continue through this lesson, keep an eye on the legends used in various maps. Some maps, such as bivariate choropleth maps, have somewhat standard legend designs. Others, such as the one used in Figure 7.1.2, are somewhat less so; they are designed and customized by the cartographer to suit the map’s data and purpose. Legend design is an important component of cartographic design in general, but is particularly important for multivariate maps.
Consider the legends you have made for your maps in lab thus far. For which map did you find designing the legend most challenging? Why?
As choropleth maps are the most popular type of univariate thematic map, it is not surprising that they are also commonly used in multivariate mapping. Most common are bivariate choropleths—choropleth maps that visualize two variables. Note that while cartographers have historically described maps of two data variables as bivariate, these maps can also be described as multivariate (more than one variable). In the context of this lesson and course, we will generally use the more comprehensive description multivariate maps.
The map in Figure 7.2.1 is an example of a bivariate choropleth distributed by the U.S. Census Bureau. It uses a hue progression (yellow to blue) to visually encode population density, and color lightness to visually encode population change. The way these symbols are combined is explained by the 3x3 box legend in the lower right.
You might notice that this map uses color hue to encode population density, which is a sequential quantitative variable—a design choice we have discouraged in previous lessons. In general, color lightness is a much better choice for encoding quantitative data. In this map, however, color lightness is already being used to map the other variable—population change. Creating multivariate maps sometimes requires bending the rules of cartographic conventions a bit so as to best represent all of your data.
Another commonly-used thematic map type for multivariate mapping is the proportional symbol map. Making these types of maps is often easier than making bivariate choropleth maps. As the main visual variable used in proportional symbol mapping is size, another variable can be added quite easily—color. The challenge lies in their interpretation: as the visual variables of size and color are quite different, this can make it challenging for the multiple variables on the map to be directly compared by readers.
Figure 7.3.1 above is a bivariate proportional symbol map that visualizes two variables: population by county (a quantitative variable, with the visual variable size) and coastline vs. interior (a qualitative variable, with the visual variable color hue).
Imagine you were tasked to create the map above, but instead of symbolizing points as coastline vs. interior, you were asked to symbolize all points by income per capita (in addition to population). What would you change about this map design to fit that new data?
Another method of multivariate map design is to stack multiple layers so they can be viewed simultaneously. Often, this is done by displaying proportional or graduated symbols on top of a choropleth or isoline map. An example is shown in Figure 7.3.2.
In the map above, visual emphasis is placed on the proportional symbols: they use size to symbolize a primary variable of interest—the estimated count of people in each city who arrived there after visiting a country on the CDC’s Zika travel advisory list. Another variable, Ae. aegypti (a mosquito capable of transporting the Zika virus) abundance, is visualized with color lightness/hue. A third variable—the approximate observed maximum extent of this mosquito, is visualized in the background for additional context. Note the careful legend design.
Making a map such as this one is a challenge, but is an example of how related variables can be mapped together to create an engaging and useful map.
Nelson, Elisabeth S. 1999. “Using Selective Attention Theory to Design Bivariate Point Symbols.” Cartographic Perspectives Winter (32): 6–28.
Thus far, we have discussed several methods for visually encoding maps with multiple variables via the addition of map symbols. There is another popular option: encoding data by altering the map’s shape or size itself. Area cartograms are maps in which the areal relationships of enumeration units are distorted based on a data attribute (e.g., the size of states on a map might be drawn proportional to their populations) (Slocum et al. 2009).
Figure 7.4.1 shows a choropleth map of Social Capital Index ratings (Lee 2018) at the top, and two cartograms beneath it. Each of these maps encode every state's Social Capital Index ranking using a multi-hue sequential color scheme. The bottom two cartograms also distort the area of each state by sizing them based on their population—but they use different techniques for doing so.
In Figure 7.4.1, the map on the bottom left is a density-equalizing, or contiguous cartogram. Though areas are distorted, connections between the areal units (here, states) are maintained. The map on the right, conversely, is a noncontiguous cartogram. States are still sized according to their population, but this method used does not require the maintenance of connections at areal boundaries. The relaxation of this requirement allows areas to be re-sized without their shapes being particularly distorted. The inclusion of state political boundaries on this map also allows the reader to make an interesting comparison: which states are disproportionally populated, and which as disproportionally less so?
Think back to earlier lessons—how might you apply color differently to improve the maps in Figure 7.4.1?
An alternative technique to constructing cartograms, called “Value-by-Alpha” mapping, was recently defined by Roth, Woodruff, and Johnson (2010). Rather than re-sizing areas based on their population, value-by-alpha maps use transparency to fade less-populated areas into the background, giving areas of higher population greater visual prominence. Thus, they serve a similar purpose to cartograms, but do not distort the map’s geography. This is not to say that they should always be used instead of cartograms—but they are perhaps an appropriate alternative when the shock value of a cartogram is undesirable, and maintenance of both area borders and shapes is desired (Roth, Woodruff, and Johnson 2010), which is not possible with traditional cartogram maps.
The examples we have explored so far have only visualized two or three variables at once. Occasionally, you may want to visualize more. One possible solution is to design data graphics that can then be incorporated into your map. A classic example of this is the use of pie charts as proportional symbols: an example is shown in Figure 7.5.1 below.
A more recent (and more complicated) example is shown in Figure 7.5.2.
Though the use of overlay glyphs does permit the addition of many variables onto the map, this does not mean it is always the best solution. As shown in the above examples, including a large amount of data in a map can make it challenging to interpret. Additionally, multivariate glyphs in general—and pie charts in particular—have well-documented disadvantages in terms of reader comprehension (Tufte 2001). Adding graphics that are already challenging for people to understand to maps tends to exacerbate such issues. This is not to say that they should never be used, however—just with caution. And fortunately, there are ways in which such maps can be made easier to interpret.
One way that multivariate maps can be made more comprehensible is through the addition of user interaction. Figure 7.5.3, for example, is challenging to interpret as a static image, particularly as the glyphs used are quite small. However, this is an interactive map. Clicking on a state creates a more informative pop-up, shown in Figure 7.5.4.
We will discuss the merits and challenges of map interactivity further in Lesson 8.
Explore the use of multivariate glyphs to explore data about well-being [186]. Can you think of ways in which this data might be symbolized instead as a static map or maps?
Despite the difficulty of creating maps with multivariate glyphs, cartographers have long attempted to tackle this challenge. One particularly interesting example of this is Chernoff faces. Chernoff faces are glyphs created by mapping variables onto facial attributes. When mapping the variable average household income, for example, a bigger smile might indicate a higher income level.
The Chernoff face technique was first proposed by Herman Chernoff in 1973. Chernoff's intention was to capitalize on the ability of humans to intuitively interpret differences in facial characteristics—both by subconsciously noting important differences in expressions that are almost unmeasurable—and by being able to ignore large differences when these differences are not relevant in context (Chernoff 1973). Chernoff also noted that his method was desirable as it permitted the designer to map many variables (as many as 18!) onto just one graphic.
Chernoff’s original application of his technique used fossil and geological data, but Chernoff mapping is more commonly used to depict social data such as well-being, or other topics related to the emotions that might be intuitively encoded using facial attribute variables. The history of Chernoff mapping is rife with controversy—some Chernoff maps such as this one: Life in Los Angeles by Eugene Turner, 1977 [187], have been heavily criticized for their use of stereotypical facial attributes and a cartoonish over-simplification of complex issues.
In response to these critiques, some cartographers have developed techniques for utilizing the advantages of Chernoff faces without some of the downsides. Heather Rosenfeld and her colleagues, for example, proposed using Zombieface glyphs rather than human faces—maintaining the emotive content and still capitalizing on people's ability to intuitively interpret facial features, but removing the human context and thus lowering the likelihood of reinforcing harmful stereotypes (Figure 7.5.6).
Take a closer look at the legend of this map—which demonstrates how the hazardous waste data was mapped to Zombie facial attributes—in the image below. As you can see, the map focuses on visualizing the presence of unknowns and uncertainty in the mapped dataset. We'll discuss further techniques for visualizing uncertainty later in this lesson.
Chernoff Zombies are among several creative solutions recently proposed: a fun example is shown in the following quasi-Chernoff map: Mapping Happiness [190]. It maps happiness, or well-being, across the United States using emoticons. Though these abstract icons cannot easily encode as many data attributes as Chernoff faces, they share the benefit of visualizing data at-a-glance using facial expressions.
Esri Blog: Chernoff Faces [191] by John Nelson.
As demonstrated by previous examples, multivariate maps are often challenging—both for cartographers to create and for readers to interpret. The term multivariate map is typically defined as a map that displays two or more variables at once (Field 2018). There is another popular option, however—comparing multiple maps. One common and often useful technique is to design a layout with many maps and show them in a progression—this is called small multiple mapping.
Small multiple maps are particularly useful for depicting data over time, as they can be arranged in a linear sequence, the way that time is typically depicted. With the increasing popularity of web-maps, small multiple mapping is often replaced with an animated map—in such maps, each map shown is shown as an individual time-stamped frame. Despite the advantages of animated maps (e.g., creating visual interest, efficient use of layout space), there are still benefits to traditional small multiple mapping. One primary advantage? The ability to simultaneously compare the various maps.
While multivariate maps are often engaging and visually interesting, it is important to keep in mind the alternatives available. We can imagine combing the set of maps in Figure 7.6.1 with some sort of transparent layering, or perhaps with an animated map that alternates between the two views. In this case, however, such designs would likely add little to the presentation value of these maps and be quite challenging to create. Here, simple works well.
So far, we have discussed two ways of mapping multiple variables—combining visual variables to encode multiple variables into one map, and visually comparing sets of maps of different data. There is a third, considerably different method that is often used for mapping multivariate data sets: cluster analysis. Cluster analysis refers to mathematical methods used to combine multiple quantitative variables into one map (Slocum et al. 2009).
There are multiple methods for clustering, the most popular of which is the K-Means algorithm, the goal of which is to identify groups of like observations based on several attributes—groups are assigned in a way that minimizes intra-group differences, while maximizing inter-group differences. Consider, for example, that you are interested in visualizing education, income, and access to green space in the US by county. You could map these three variables individually, or you could use cluster analysis to identify groups of counties that are similar along all three dimensions. Once such groups are determined, you could map them with a qualitative color scheme onto a chorochromatic map.
Cluster analysis is a complicated topic, and we will not go into its details in this course. What is important to understand is that it provides a mathematical alternative to the other more design-based multivariate mapping techniques we have explored so far. You are encouraged to explore the recommended readings if you are interested in learning more about cluster analysis and about implementing it in GIS.
ArcGIS Pro Tool Reference: How Multivariate Clustering Works [195]. Esri 2018.
Jain, A. K. 2009. "Data Clustering: 50 years beyond K-Means [196]." Pattern Recognition Letters.
Of the many variables you may wish to include in your maps, there is one that has received particular focus from cartographers due to its unique characteristics—uncertainty. Uncertainty is a complex concept which has been defined differently by various authors. For example, Longley et al. (2005) define uncertainty as "the difference between a real geographic phenomenon and the user’s understanding of the geographic phenomenon." We use this definition as it encompasses the many variations of uncertainty that in emerge during multiple stages of map-making—during data collection, data classification, visualization, map-reader interpretation, and more (Kinkeldey and Senaratne, 2018).
It can be assumed that all geographic data contain some level of uncertainty. A map of average income by county, for example, might classify a county as having an average household income between $50,000 and $60,000. Despite this, it is possible that the actual value falls outside of this range—due to survey response errors, non-response to survey (e.g., Census) requests by some residents, or changes in the data over time (e.g., some survey respondents have moved in or out of the county since the data was collected). A map of precipitation levels, similarly, will also contain uncertainty, likely due to the imprecision or inaccuracy of measurement instruments, but possibly due to human error or related factors as well.
A helpful list of terms and definitions related to uncertainty can be found here: Kinkeldey, C., & Senaratne, H. (2018). Representing Uncertainty [197]. The Geographic Information Science & Technology Body of Knowledge (2nd Quarter 2018 Edition), John P. Wilson (ed.). DOI: 10.22224/gistbok/2018.2.3
Traditionally, researchers have grouped geodata uncertainty into three categories – the what (attribute/thematic uncertainty), the where (positional or locational uncertainty), and the when (temporal uncertainty) (MacEachren et al. 2005). The success of visual variables for depicting uncertainty depends on the type of uncertainty to be mapping. Containing a point within a colored glyph or circle, such as Google’s “blue dot,” might be most effective for depicting positional uncertainty (Google Maps; McKenzie et al. 2016). Use of another variable such as transparency might be more effective for depicting attribute uncertainty, such as uncertainty of unemployment rates in a county-level map.
Like other multivariate data, uncertainty can be combined with the other visualized data in a map, or compared by visualizing it in a separate map view. Figure 7.8.1 shows two maps that use different techniques to visualize the uncertainty in the data. Figure 7.8.1 (top) uses a combining technique, in which a visual overlay is used to show attributional uncertainty. Figure 7.8.1 (bottom) uses a reliability diagram—an inset map that the reader can reference to understand which locations on the map contain the most certain data values. In general, the combining method is a more popular technique, though a compare technique might be useful if the primary map is sufficiently complex, and thus adding overlay would make the map difficult to comprehend.
Among combined uncertainty visualization techniques, methods for visualizing uncertainty are typically classified as either intrinsic or extrinsic. Intrinsic uncertainty visualization techniques cannot be visually separated from the visualization of one or more other variables, while extrinsic visualization techniques are easier to interpret separately. An example of the difference between these two techniques is shown in Figure 7.8.2.
In Figure 7.8.2, both extrinsic (top) and intrinsic (bottom) uncertainly visualization techniques are shown. The extrinsic visualization uses a hatched fill overlay to denote uncertain values—thus, the visualization of uncertainty is visually separable from the visualization of the data underneath. Figure 7.8.2 (bottom) by contrast, uses an intrinsic visual variable—transparency—to visualize data uncertainty. The two variables are combined together to create the legend as well.
Any visual variable can be adapted to demonstrate uncertainty. However, some have been developed specifically for this purpose. MacEachren (1995) proposed the idea of clarity as a visual variable for static maps, an overarching concept that can be further divided into three visual variables: transparency, crispness, and resolution (MacEachren 1995). Transparency is a somewhat familiar visual variable, as it has been adapted for other purposes than displaying uncertainty, such as in the value-by-alpha maps described earlier in this lesson.
Crispness is a particularly intuitive way of visualizing uncertainty. Features are depicted on a continuum from crisp to blurry, with less certain values appearing appropriately out-of-focus (Figure 7.8.3).
Resolution creates a similar effect—features with less certain boundaries or attributes are depicted in courser resolution, suggesting a lack of certainty in the map.
These visual variables are popular for depicting uncertainty as they intuitively suggest uncertainty (or certainty) by design. Just as higher data values are visually encoded with larger symbols, less certain boundaries, for example, may be visually encoded with fuzzy boundaries.
Though uncertainty is often discussed in terms of uncertainty within data due to imprecise instruments, imperfect collection methods, etc., an important additional context where uncertainty plays a role is in the mapping of future scenarios. Climate models, for example, use past and present data to predict future conditions, but these predictions are inherently uncertain. Figure 7.8.5 below contains maps of temperature and precipitation change predictions. The first map (top left) maps the average result of 37 predictive models intended to estimate temperature change by 2050 (Kennedy 2014). The middle map shows the warmest 20% of models—the 20% coldest models are summarized at the right. The bottom three maps show a similar comparison of maps created from precipitation models.
Unlike previous examples, these maps do not use intuitive visual depictions of uncertainty. However, the map-maker's inclusion of all three maps for each data variable shows the range of possibilities that might lie ahead: the future is always an uncertain entity. It is implied that these maps depict not all possible scenarios but a range of likely ones; they intend not to precisely predict the future but to help users understand what might future conditions they might expect to come about.
In the last section, we discussed how to conceptualize uncertainty, and ways in which it can be visualized. One important question remains: why should we do so? Creating well-designed maps can be challenging, and adding a depiction of uncertainty makes this process even more so.
Uncertainty is typically depicted in maps for two primary reasons: (1) its inclusion may be regarded as an ethical necessity—many maps are created with significantly uncertain data, and a cartographer might feel that withholding this information from the map reader would be misleading. (2) Consideration of uncertainty plays an important role in decision-making, and thus its visualization might be necessary in some contexts—for example, maps of predictive hurricane paths tend to include a “cone of uncertainty” (Figure 7.9.1)—and such maps often play an important role in decisions made by residents of storm-affected areas.
So how does the visualization of uncertainty effect decision-making with maps? Kinkeldey et al. (2015) conducted a review of studies that attempted to answer this question. Most of the studies they analyzed suggested that the visualization of uncertainty does have an effect on task performance with maps and similar spatial displays (Kinkeldey et al. 2015). Simpson et al. (2006), for example, studied the use of uncertainty visualization in surgical tasks with graphic displays, and noted that the inclusion of uncertainty visualization improved performance accuracy. The positive influence of uncertainty visualization on task-completion accuracy with maps is a somewhat common finding. Though findings are less consistent with regards to task completion times (i.e., speed), uncertainty visualization seems at least not to significantly increase task-completion times (Kinkeldey et al. 2015).
Despite this, there is still not a consensus concerning whether uncertainty visualization is always helpful for decision-makers—some studies note that participants perceive uncertain data as risky, which can induce irrational decision-making via loss-aversion (Hope and Hunter 2007). Whether uncertainty visualization is useful—and whether it is useful enough to warrant the design efforts it requires—is context dependent and still thoroughly up for debate.
For Critique #4, you will be reviewing a colleague's map from Lab 6: Terrain and Trails Visualization. Lab 6 focused on terrain visualization, as well as the symbolization of overlay data to create a map for the imagined Paradise Valley Trail run in San Francisco, California. While completing this critique, you should attempt to view your classmate's map from the perspective of its intended reader - a registered trail runner or one of their supporters. In other words, how successful do you think the map would be, for example, assisting you navigating to the race or determining how best to train or compete in the race?
For this assignment, write a 300+ word critique of your classmate’s Lesson 6 Lab (as assigned).
In your written critique please describe:
Your map critique should be constructive and, as suggested above, should focus as much on what the map does well as it does on suggestions for improvement. Due to this lab's specific focus on map audience and purpose, you may find it helpful to reflect upon for whom (e.g., runners vs. spectators) elements of the map's design might be most helpful. You should connect your ideas back to concepts we have discussed in the course content thus far.
Please list the student name of the map you have been assigned at the top of the page.
A rubric is posted for your review.
You will work on Critique #4 during Lesson 7 and submit it at the end of Lesson 7.
Step 1: Upon notification of the Peer Review (Critique), go to Lesson 6: Lab 6 Assignment. You will see your assignment to peer review. (Note: You will be notified that you have a peer review in the Recent Activity Stream and the To-Do list. Once peer reviews are assigned, you will also be notified via email.)
Step 2: Download/view your classmate's Lab.
Step 3: Write up your critique using the prompts above in a Word document. Be sure to also review the rubric in which you will be graded for Critique #4 for more guidance. Save your Word document as a PDF. Use the naming convention outlined below.
Step 4: In order to complete the Peer Review/Critique, you must
- Add the PDF as an attachement in the comment sidebar in the assignment.
- Include a comment such as "here is my critique" in the comment area.
- PLEASE DO NOT use the rubric in the lesson assignment to award points or grade the map you are critiquing. Only submit your PDF document.
Step 5: When you're finished, click the Save Comment button. You may need to refresh your browser to see that you've completed the required steps for the peer review.
Note: Again, you will not submit anything for a letter grade or provide comments in the lesson rubric.
For this lab, imagine you are the mapping specialist and a member of a new “Happier World 2020” team, tasked with creating a report of maps and supporting text (500+ words) to present to a group of leaders at the United Nations. The total length of the report should be no longer than 7 pages (this page count is inclusive of all text and maps).
Use data from the 2018 World Happiness Report [201] to visualize multiple variables related to world happiness and well-being; these relate to the UN’s Sustainable Development Goals for 2030 [202]. In creating this document, you will gain experience both with creating maps from a multivariate data set and in visualizing data uncertainty. In compiling these maps as a document rather than in a map layout, you will explore a new way to present your maps - one that is common when designing maps to present to policy-makers, or when illustrating a scientific paper.
For Lab 7, your only deliverable will be a single compiled PDF with text and images of your four maps.
Export your main map and three smaller maps as images (or use the Snipping tool) and include them as figures in a report with accompanying text. As when creating a map layout, attend to aesthetics, visual hierarchy, and negative space.
A rubric is posted for your review.
Please refer to the Lesson 7 Lab Visual Guide.
You can learn more about these indicators in the World Happiness Report [201], but this will help you get started as you decide what indicators to map and what ideas you might propose in your report.
The Life (Cantril) Ladder asks people to imagine a ladder with steps from 0 to 10 (the top), where the top of the ladder is their best possible life, and the bottom is the worst—and to place their life at the present at some point on the ladder.
GDP per capita is calculated in terms of Purchasing Power Parity (PPP) in 2011 international dollars as indicated by the World Bank.
Healthy life expectancy is calculated based on data from the World Health Organization (WHO) and WDI. WHO data is based on estimates from 2012.
Social support is the national average of responses (0 or 1) to the question “If you were in trouble, do you have relatives or friends you can count on to help you whenever you need them, or not?” This question is part of the Gallup World Poll (GWP).
Freedom is calculated as the national average of responses to the question “Are you satisfied or dissatisfied with your freedom to choose what you do with your life?” This question is part of the Gallup World Poll (GWP).
Generosity is calculated as the residual of regressing the national average of answers to: “Have you donated money to a charity in the past month?” on GDP per capita. This question is part of the Gallup World Poll (GWP).
Perceptions of corruption is calculated as the average of answers to two questions: “Is corruption widespread throughout the government or not?” and “Is corruption widespread within businesses or not?” Perception of business corruption is used as a proxy for a total corruption measure in countries where responses to questions of government corruption are not available. These questions are part of the Gallup World Poll (GWP).
Helliwell, J., Layard, R., & Sachs, J. (2018). World Happiness Report 2018, New York: Sustainable Development Solutions Network.
Credit for all screenshots is to Cary Anderson, Penn State University; Data Source: World Happiness Report (Helliwell et al. 2018), Natural Earth.
This is your starting file in ArcGIS Pro: It contains boundary data from Natural Earth, as well as thematic data from the World Happiness Report [205].
As suggested above, the purpose of this lab is to visualize data from the World Happiness Report. You should explore this data in ArcGIS Pro (Figure 7.2), and view the World Happiness Report [205] online to get a sense of what the data means. Definitions for each happiness indicator variable are also listed at the bottom of the Lab 7 requirements page.
Your first task is to choose a happiness variable of interest and to map this as your primary map for the lab. For this map, you must select one of the variables with an associated "Standard Error" column. We will be using this as a proxy for uncertainty (more on that later).
You may choose from among several thematic mapping options (choropleth; graduated symbol; proportional symbol, etc.) to map your happiness data. This survey data is a bit more abstract than the data we have become accustomed to working with, so for this lab, you have a bit more freedom than usual in selecting a mapping method. Shown below are some examples of symbolization methods that you might use for your primary map.
The data provided to you for this lab has already been standardized for you, so you will not need to choose a "normalization" field.
When designing your map symbols, recall previous labs and focus on creating a useful and aesthetically-pleasing design. If using proportional symbols, for example, you will likely want them to be semi-transparent so that both symbols can be seen in the case of overlap.
As you might notice, not every country is included in the World Happiness Report. Due to this, an important element of your map design will be deciding how to visualize null values. You want it to be clear to the map reader which countries are not included in the report, but you do not want these to be too prominent in your map's visual hierarchy - they should not distract from your map's main purpose. As shown below, symbolizing null or "out of range" values is a slightly different process depending on which symbolization method you choose.
Your primary map should visualize not only the happiness indicator you have selected, but also its associated uncertainty. For this lab, you will use your chosen variable's associated "standard error" field as a proxy for uncertainty. Assume that higher standard error = higher uncertainty. Though the statistics involved are slightly more complicated than this, the focus of this lab is on visualization and thus this generous assumption is suitable for our purposes. Additionally, as we are interested in where the data is more or less certain, rather than in the actual standard error values, you should classify this uncertainty data into general groupings (e.g., "low," "medium," "high").
You may choose to visualize uncertainty either extrinsically or intrinsically. When selecting a method, consider how you will represent this uncertainty in your map's legend, as well as how your design might be interpreted by your map's intended audience. This guide presents two popular methods for visualizing uncertainty, though there may be others.
Option #1: Extrinsic uncertainty visualization
The easiest way to create an extrinsic uncertainty visualization layer is to copy-and-paste your country-boundary layer (which includes all the happiness data), and to symbolize uncertainty with the duplicate layer.
Two examples of extrinsic uncertainty visualization are shown in Figure 7.9. Recall from the lesson content the visual variables most effective for visualizing uncertainty. Your goal should be to create an intuitive design.
Once you've finished creating your map and adding it to a layout, you'll need to design an import part of this lab - your map legend. Figure 7.10 below contains some examples of legend designs. More so than with previous labs, you will likely want to make significant edits to your legends via the "convert to graphics" function in ArcGIS Pro.
Option #2: Intrinsic uncertainty visualization.
The other primary option for visualizing uncertainty in your map is intrinsically, via the "vary symbology by attribute" option in ArcGIS Pro. Think carefully about how you apply elements like transparency—the progression of your data should match the visual progression of your design. To accurately depict your data and its uncertainty, you may have to manually edit or reverse a color scheme or transparency range.
Once you've finished your primary map, you will create three smaller maps. These maps should visualize three additional variables from the World Happiness Report. You do not need to visualize uncertainty in these smaller maps. For this lab's final deliverable, you will include your four maps in a report that focuses on answering the following questions:
You may start with this template: Report Template: Lab 7 [206] and customize it as you wish. You do not need to include additional research from sources outside of the World Happiness Report, though you may if you wish.
Remember design ideas from previous labs - you may want to add elements such as a grid, explanatory text, and data credits to your layouts. It's up to you how you balance your final document with such elements - for example, instead of listing a data source on your map images, you can simply include this source in the text you write. Also, note that due to the scale of these maps you do not need a north arrow or scale bar - focus on creating a useful and cohesive document, as well as smartly-designed legends.
You’ve reached the end of Lesson 7! In this lesson, we explored the concept of multivariate mapping, wherein multiple data attributes are visualized in one map. We discussed the many ways in which cartographers visualize complex data, including with multivariate choropleths, cartograms, cluster analysis, and multivariate glyphs. In the case of a particular additional variable—uncertainty—we discussed intrinsic vs. extrinsic uncertainty depiction, as well as usefully intuitive visual variables such as crispness and resolution.
In Lab 7, we compiled a document with a set of maps intended to influence a group of decision-makers at the United Nations. In doing so, we explored the challenge of mapping multivariate data (and used the compare small-multiples technique), as well as adequately visualizing uncertainty. In this course, we have often made multiple maps in one lab—this assignment provides an example of how such maps might be combined with text in a practical and important mapping task. In Lab 8, we take on a less serious topic—creating an interactive basemap inspired by a favorite work of art.
You have reached the end of Lesson 7! Double-check the to-do list on the Lesson 7 Overview page [207] to make sure you have completed all of the activities listed there before you begin Lesson 8.
Welcome to Lesson 8! In this lesson, we discuss another key topic in cartography: map generalization. Generalization, or transforming a map’s features (traditionally from large-scale to small-scale) to fit a map’s given scale and purpose, has been increasingly in focus given the proliferation of interactive multi-scale web maps. Such maps are among a new generation of 'fast maps', which include interactive, animated, and viral maps and mapping products. These maps, as well as other products of technological advancements in mapping, including 3D maps and extended reality applications, present new challenges and opportunities for geographers across many fields of interest.
In Lab 8, we break from our focus on mapping in ArcGIS to design an interactive multiscale basemap using the online map design platform Mapbox Studio. To highlight the creative designs possible with such tools, we design this map by taking inspiration from a favorite piece of media/art. Lab 8 thus ties together our discussion of generalization and interactivity with previous discussions of maps for advertising, map symbology, and basemap design.
Action |
Assignment | Directions |
---|---|---|
To Read |
In addition to reading all of the required materials here on the course website, before you begin working through this lesson, please read the following required readings:
Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allow. |
The required reading material is available in the Lesson 8 module. |
To Do |
|
|
If you have questions, please feel free to post them to Lesson 8 Discussion Forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
In Lesson 7, we discussed uncertainty visualization—yet another component of a common theme in cartography: transferring the complexities of our world into a visualization via a map. When we transform the many features of Earth’s geography into a form more appropriate for a map’s given scale and purpose, this is called cartographic generalization. Thorough understanding of generalization and the related concept of scale is—and has always been—essential for creating high quality maps. The increased prevalence of web-maps, which permit zooming and panning across multiple extents and scales, has encouraged increased research in these topics. In this lesson, we discuss generalization, both in general, and in the context of multi-scale and interactive web maps.
All maps contain some level of generalization—maps would be unusable otherwise. Representing every element of the real world on a map is not feasible, nor would such a map be interpretable by readers. Generalization permits cartographers to construct maps with an appropriate level of detail. In Lesson 6, we discussed the necessity of using the correct resolution of (raster) digital elevation data to create terrain visualizations. In this lesson, we focus primarily on the generalization of vector data, such as hydrologic features and political boundaries.
When considering what level of detail is appropriate, it is important to consider your map's location, scale, and geographic extent. A map of seaside hotel locations in Massachusetts would, for example, show a much more detailed coastline of Cape Cod than would a map of the entire United States.
Natural Earth [208] is a popular source of boundary data that we have used extensively in this course. Figure 8.1.1 below demonstrates the differences in level of detail between different boundary datasets that Natural Earth offers. The purple boundaries (left) show the most detail. Such data are appropriate for maps of large regions (e.g., scale = 1:10m). The pink (center) boundaries would be better suited for small-scale maps of continents or the entire globe (e.g., scale = 1:50m). The blue (right) boundaries are heavily generalized, and would be best suited for very small-scale maps, or maps meant to emphasize style over accuracy (e.g., scale = 1:110).
Figure 8.1.2 shows each of the above boundary files at an appropriate scale given their level of detail. The extent of the largest-scale map (left) is shown by the extent indicator in the center and right maps.
For an interactive experience with generalization, try uploading a shapefile from NaturalEarth [209] to the interactive tool MapShaper [210].
So far, we have talked about the overall idea of generalization – using data that is the correct level of detail for your map’s scale. A general-purpose map of a small town, for example, would likely show lakes, ponds, and reservoirs, while a small-scale map of a large region would show only the largest waterbodies (e.g., rivers, large lakes, and oceans). Often, objective rules are used to determine what elements are displayed on a map (e.g., “only show lakes that cover more than five square miles”). However, due to the uneven distribution of features across the landscape, cartographers also have to make some generalization decisions that are complex, subjective, and specific.
An example of this is demonstrated by Figure 8.1.3. Some cities are labeled, and some are not. At first, it may appear that the largest cities are labeled, and to some extent this is true. New York, NY is labeled, as well as Washington, DC. However, you may notice some cities that are absent—most notably Philadelphia, PA. A city with 1.5 million people is left off the map, while Reading, PA—a city of about 88,000—is included. Why?
Philadelphia is located in a densely-populated region, with many nearby cities, such as Trenton, Baltimore, and Washington, D.C. By contrast, Reading, PA is surrounded only by smaller towns. Web-maps are designed to display—or not display—city labels based on a number of factors. These include population and general importance, but also design-relevant factors, such as the density of labels on the map.
OpenStreetMap (figure 8.1.3) is designed as a general-purpose map; the maps you create will typically have a more specialized purpose. And if your map’s topic was related to the city of Philadelphia, you would be sure to use your judgment to adjust the decisions made by OpenStreetMap’s generalization algorithm.
Chapter 3: Map Generalization: Little White Lies and Lots of Them. Monmonier, Mark. 2018. How to Lie with Maps. 3rd ed. The University of Chicago Press.
As suggested by the previous OpenStreetMap example, generalization is a process for dealing with conflict and congestion among map symbols—a strategy for creating a more readable and useful map. Though this is a complex and context-dependent problem, some resources are available to help you determine the appropriate level of detail for your maps. ScaleMaster (Brewer et al. 2007) for example, available at scalemaster.org (Note: As of July 28, 2023, the scalemaster.org site no longer exists), offers advice to mapmakers on which features ought to be included at different scales, and for different mapping purposes.
We will not go into the details of ScaleMaster in this lesson, but you are encouraged to read more about this idea through the Cartographic Perspectives article if you are interested. The most important takeaway is that different scales require differing levels of detail, and that the appropriate level of detail is mediated by the map’s context (e.g., topographic vs. zoning maps).
Generalization can be broadly categorized as either selection or symbolization. Selection is simple—it refers to the decision of whether to include (or not include) a feature at a certain scale. Symbolization refers to alteration of the way a feature is designed in order to make its design more appropriate for the scale at hand. For example, when designing a small-scale map, you might choose to not include cities unless they are high population (selection), and to symbolize these cites as labeled points rather than as areas (symbolization). Generalization traditionally refers to reducing detail in a map as much as is necessary to maintain legibility and usefulness at a specified scale. Generalizing multi-scale web maps (which exist at many rather than one scale) is more challenging, but not fundamentally different—we can think of every possible scale step (or zoom level) of a multi-scale web map as its own map for which an appropriate level of detail must be determined.
As generalization is a fundamental topic in cartography, many cartographers have proposed theoretical frameworks for discussing generalization. For simplicity, in this lesson, we will focus on the set of generalization operators recently proposed by Roth et al. (2011), as they were developed based on a comprehensive review of previous literature. As we discuss generalization operators, an important distinction should be made between generalization operators and generalization algorithms. Operator refers to a cartographer’s conceptualization of an intended change (e.g., I want to remove some roads to reduce the visual clutter of this road network), while an algorithm is a system followed for implementing this idea (e.g., I will remove all roads with speed limits below 25mph) (Roth et al. 2011). Like Roth et al., we focus on operators rather than algorithms in this lesson as they are more widely applicable to map-making tasks, and not dependent on the use of specific datasets or GIS software tools.
Roth et al. (2011) classify feature generalization operators into three groups: content, geometry, symbol. Content operators directly alter the content of the map, typically by adding or removing features at particular scales. An example would be deciding not to include local roads or trails in a small-scale map. These operators include: add, eliminate, reorder, and reclassify.
Geometry operators describe the ways in which different features' geometry can be altered to create a map that is more legible and aesthetically pleasing. Examples include smoothing a line feature and representing a city as a point rather than an area. Geometry operators include: simplify, aggregate, collapse, merge, displace, exaggerate, and smooth.
Symbol operators alter feature symbology to improve legibility, but do not change the features’ underlying geometry. An example would be simplifying the pattern in an area fill so it still looks good at a smaller scale. Symbol operators include adjust color, enhance, adjust iconicity, adjust pattern, rotate, adjust shape, adjust size, adjust transparency, and typify.
It is not necessary to memorize the above operators, but you should aim to understand the difference between the three groups of operators (i.e., content, geometry, symbol) and think critically about situations in which each might be useful.
The advent of the world wide web initiated many changes in the world of map-making. Though centuries-old cartographic principles are still relevant in a web-mapping world, digital map-making has presented new unique opportunities and challenges for cartographers.
The increasing ubiquity of the Internet has influenced cartography in many ways, from changing the nature of maps themselves (e.g., with new interactive and animated maps), to facilitating a system wherein map-making tools are widely accessible—a world in which almost anyone can make and widely-share a map.
Figure 8.3.1 demonstrates the evolution of the popular GIS software ArcGIS, from ArcMap/ArcView to the newly-released ArcGIS Pro, designed with modern graphics, searchable toolboxes, and a ribbon-based interface. Perhaps even more indicative of the times is the widespread availability of web-based mapping tools and libraries, including CARTO [215], Leaflet [216], Mapbox [217], Social Explorer [218], and many more.
Geographer Mark Monmonier (2018) uses the term “fast maps” as an umbrella term to describe many new forms of maps and mapping products that have come about in the internet age. These include interactive maps, animated maps, and viral maps—maps that may be static or otherwise but are nevertheless a product of new technologies and widely spread due to the Internet and social media. New interests in virtual and augmented reality have also added to the variety of maps available in this widely-connected world.
If you have several years of experience using GIS Software, consider how this software has changed over the course of your career. What software did you use when you were first learning GIS? How is it different from ArcGIS Pro?
Chapter 14: Fast Maps: Animated, Interactive, or Mobile. Monmonier, Mark. 2018. How to Lie with Maps. 3rd ed. The University of Chicago Press.
We briefly discussed interactive maps in the previous lesson on multivariate mapping—interactivity is often used to solve problems related to multivariate mapping, such as the challenge of fitting all the necessary data into one map frame. New technologies (most notably, mobile smart phones) have both increased the challenge of designing maps and contributed their own solutions. Creating a map that can be viewed on a 4.7-inch screen, for example, can be quite a difficult design problem. Yet, accessibility to mobile cell data and location-aware devices have enabled the creation of zoomable, pan-able, user specific maps—thus reducing the amount of map content required in-view at any one time.
Web maps such as the one in Figure 8.4.1, which allow the user to zoom and pan around the extent of the map, are commonly called slippy maps. While they may serve as general purpose maps themselves, these maps are most often used as basemaps that provide location context for a variety of thematic or functional overlays—such as the traffic volume data or navigational functionality of Google Maps.
We often categorize interactive maps by the level of user interaction (low to high) they permit. Some maps allow only simple interactions such as panning or zooming, or perhaps show additional information about features on mouse hover or click. Others may be developed for expert users, and include the ability to search, filter, and analyze data, as well as the option to upload the user’s own data for exploration and analysis. Many interactive maps, such the one in Figure 8.4.2, fall somewhere in the middle of this continuum.
While the term interactive map is most often used to describe maps such as the one in Figure 8.4.2, other maps are better characterized as containing passive interactivity. These are maps that respond to actions of the user, though not in the traditional sense of a user interacting with tools via a map interface (Monmonier 2018). An example of this is an automotive personal navigation device (PND).
Though these devices also contain traditionally interactive components, they are primarily designed to respond to one particular user behavior—movement through space and time. Such mapping tools provide navigation, real-time traffic, and safety-zone warnings; some even provide advanced notifications such as lane departure warnings via unit-mounted cameras and other sensors.
Interactive maps have the potential to be useful in any geographic decision-making context, wherein the map can provide an appropriate interface between the human and the machine (Monmonier 2018). Due to the complexity of many of these products, however, the effectiveness of an interactive map is often dependent not only on the design of the map itself, but on its interface and related functions. This has made interactive map-making a particularly interdisciplinary subset of cartography, as successful approaches borrow increasingly from research in data visualization, human-computer interaction (HCI), and computer science. We will discuss the interface between maps and their users in more detail in Lesson 9.
Roth, Robert E. 2013. “Interactive Maps: What We Know and What We Need to Know.” Journal of Spatial Information Science 6: 59–115. doi:10.5311/JOSIS.2013.6.105.
Though animated maps may also have interactive components, they are uniquely defined by their use of animation to display spatial data. A type of animated map you have likely seen and used is a weather radar map, such as the one shown in Figure 8.5.1. These maps typically contain little user-interaction capabilities—they are watched by the user as if watching a movie—though they may contain zooming or panning functionality, or the option to pause the animation at a point in time.
Animated maps are used to visualize a wide range of data topics, from weather to health data, demographic statistics to travel routes. Most common among these maps is the inclusion of time as the variable that is changed as the animation is performed. Though, theoretically, any quantitative variable could be depicted via animation, the use of animation to depict data through time is supported by the congruence principle which states that the external graphic representation of data should match its intrinsic characters (e.g., in the case of animation, the animation plays across time, and represents temporal data) (Tversky, Morrison, and Betrancourt 2002).
Despite the popularity of animated maps for data visualization, little research has yet been conducted that supports its use as a replacement for static graphics such as small multiple maps (Tversky, Morrison, and Betrancourt 2002; Griffin et al. 2006). Animated maps present unique challenges for users, who are often hindered by perceptive constraints, such as change blindness – the inability to detect changes in maps across animated frames, often combined with user overconfidence in map comprehension (Fish, Goldsberry, and Battersby 2011).
Griffin et al. (2006) conducted a map-cluster detection study with animated maps and small multiples and found that users did tend to be more successful with animated rather than static maps for this task. They note an important challenge in animated map design, however—that the pace or speed of the animation is influential on user success, and that different paces are more useful for different maps. There is no ideal animation pace for maps, though cartographers ought to consider what pace might be most useful for their map’s intended audience and purpose. One way to sidestep this decision is to add simple interaction features to animated maps, such as the ability to pause or step through time, so the user might adjust the animation to a speed that works best for them (Tversky, Morrison, and Betrancourt 2002). Though many visual variables are used in animated mapping, pace is among the visual variables used specifically for encoding data via animation. Other animation-relevant variables include rate of change – how much the map changes between each animated frame, and order (DiBiase et al. 1992), which is the order in which individual frames are presented (often chronologically, but not always).
Consider a mapping purpose for which you might want to create an animated map with frames in non-chronological order. Why would this design choice benefit the map user?
Alan MacEachren (1995) extended the above-mentioned visual animation variables to include display date – the starting time of a temporal sequence, frequency – the number of unique states within each unit of time (e.g., animated frames per year), and synchronization – the coincidence (or otherwise) of time series when two or more are displayed at once (e.g., snowfall and school attendance might be displayed out of sync).
Though the term “dynamic maps” implies movement within maps (i.e., animation and interaction), we discuss here a similar category of maps, as suggested by Monmonier (2018) in his categorization of “fast maps” – viral maps. Though there is no widely-accepted definition of a viral map, the term applies broadly to a map that is shared widely, and through non-traditional processes (i.e., through users sharing content with each other, rather than from a singular, popular provider) (Robinson 2018).
Maps that spread in this way tend to inspire emotion and be persuasive in nature (Monmonier 2018; Muehlenhaus 2014; Robinson 2018). Despite the heightened study of such emotive and persuasive maps due to their dispersion on social media, persuasive maps themselves are not new. Figure 8.6.1 shows a map from the Civil War, which illustrates General Winfield Scott’s plan to conquer the south. The snake illustrates a dark, emotional message.
Social networking sites such as Twitter have facilitated the spread of maps to a global audience with incredible speed. Such sites also invite the designing and sharing of persuasive maps by nearly anyone with access to the Internet—it is difficult to overstate the contrast between this new environment of online map distribution and cartography’s history of maps being made primarily by professional cartographers or those in positions of power. In many ways, we find ourselves in an exciting, dynamic, more democratic era of map-making. It is important to note, however, the challenges that have arisen in this new era. The increasing ubiquity of maps and map-making has blurred the lines between mapmakers who make mistakes and those who deliberately mislead; between personal perspectives and dangerous propaganda.
Related to the increased availability of map-making tools and online map distribution channels, web technologies have facilitated increased access to wide amount of data within the public domain. Where debate tends to ensue, however, is when such data are made more visible and accessible to everyone, such as with the creation of an engaging map. Maps printed along with an article in a local newspaper titled “The Gun Owner Next Door: What You Don’t Know About the Weapons in Your Neighborhood” provide a useful case study of such a debate. The article and accompanying maps identified gun-owners in the local area by their names and addresses. The map itself ‘went viral’ both due to people's intrigue in the data mapped, and the outrage that the discussions surrounding it incurred.
Read the article mentioned above, available here: “The Gun Owner Next Door: What You Don’t Know About the Weapons in Your Neighborhood [227].” Would you consider it ethical to map any data, as long as it is available in the public domain? If not, where do you stand on this issue? How might we decide where to draw the line?
As a Penn State Student, you have free access to the NY Times, the Wall Street Journal, and others through the Student News Readership Program. This link [228] provides instructions on how to get access.
Maps are omnipresent in political media—consider the interactive maps used extensively on news channels while reporting election results. About a month before the 2016 US Presidential election, Nate Silver (Silver 2016) posted a map with the heading “Here’s what the election map would look like if only women voted: [229]”
In addition to reaching viral status itself, the map inspired many others to create similar maps, such as what the election map would look like if only millennials/white women/people of color voted. Robinson (2018) uses Silver’s map as an instrumental example of a viral map in his recent paper, Viral Elements of Cartography. He notes that it is characteristic of viral maps to inspire the creation of others.
Though viral and persuasive maps are often discussed in tandem (e.g., Muehlenhaus 2014), viral maps need not always be persuasive or political. The map in Figure 8.6.2 below was designed by Joshua Stevens, a cartographer at NASA who despite being well-known in the data visualization community, has only a fraction of the online following of journalist Nate Silver (Silver 2018). It was the creativity and entertainment value generated by Stevens’s map which was responsible for generating its viral status.
Like Silver’s map of women voters, Stevens’s Sunsquatch map inspired the design of many others, some of which went viral themselves, such as Jerry Shannon’s Smothered and Covered map (Figure 8.6.3) which illustrated where one could watch the eclipse while eating at Waffle House.
These maps by Silver, Stevens, and Shannon highlight the usefulness of Monmonier’s classification of new-era maps facilitated by web technologies as fast rather than dynamic or interactive maps (Monmonier 2018). The speed at which these maps were shared to thousands of users certainly qualifies them as fast, though they are simple, static maps. And though these static maps do not include animation or permit user interaction, they did instigate discussion and inspire further map-making, making them interactive in their own right. Certainly, interactive and/or animated maps can also ‘go viral.’ The above examples illustrate, however, the power in pairing a simple illustrative graphic with a creative idea.
Similar to how new web-based technologies have made it easier to design interactive and animated maps, technological advancements have altered mapmaking in another, related way—enabling more realistic depictions of the real world through more accessible 3D-mapping tools and virtual/augmented reality.
Three-dimensional visualizations have long been used to create city models (e.g., Figure 8.7.1) and similar models of Earth’s terrain or built environment. As we discussed in Lesson 6, these models are useful in that they provide a realistic view of the environment, but their realism and complexity often come at a cost. For example, the oblique view inherently obstructs some of the scene (e.g., locations behind tall buildings), and physical models are typically not built to scale.
In the past, creating complex 3D digital visualizations and physical models came at a near-prohibitory cost. Yet recent increases in the computational power of mainstream computers and new software tools have reduced the time, capital, and expertise required to create three-dimensional maps. Naturally, this has encouraged cartographers to make more of them. The inclusion of 3D visuals in mapping tools has become increasingly widespread—realistic modeling of buildings can now be seen, for example, in popular mapping applications such as Apple Maps [234].
Increasing interest in and availability of 3D mapping tools has also resulted in an increased use of extruded or perspective height as a visual variable. Unlike the 3D map examples above, perspective height uses a third dimension to encode a variable distinct from the actual physical height of a feature.
An example is shown below (Figure 8.7.2). This is a choropleth map that uses a multi-hue sequential color scheme to encode the population density of the United Kingdom by postal code. In addition to color, however, another visual variable is used—perspective height. Areas with higher densities are extruded from the map, giving them increased visual emphasis. The result is a map that portrays the look of a varied terrain—only instead of actual physical terrain, it visualizes the terrain of people across the landscape.
Similar to the visual depiction of uncertainty, an important question surrounds the use of 3D visualization: is it useful? The answer, as provided by recent cartographic research, is similar: it depends. Generally, studies have found that people enjoy using 3D maps more than their 2D counterparts. These studies also typically find, however, that people perform tasks less efficiently with 3D graphics than with simpler 2D visualizations (Smallman and John 2005).
There is no scientific consensus on whether 3D visualization tends to be helpful for users, and due to the context-dependence of such a question, it is unlikely that there will be an answer anytime soon. What the current state of research suggests is that 3D visualizations should be used with caution. In contexts where seconds count (e.g., emergency management; disaster response) for example, 3D visualization tools might be a risky option. In contexts where user enjoyment is of greater priority (e.g., in a university’s campus map), it might instead be an excellent choice.
Related to 3D visualization is a new system of technologies that has quickly gained attention in recent years: extended reality. Extended reality is an umbrella term that encompasses several related technologies including virtual reality and augmented reality. You have likely seen examples of these technologies used for sports and gaming—examples include Pokémon GO, an augmented reality mobile game, and Samsung’s Oculus Rift virtual reality headset.
Even the yellow first down line you see on the football field during NFL games is an example of augmented reality. Watch the video about it: How the NFL's magic yellow line works [237]. Can you think of an environmental or emergency management scenario in which similar technology might be useful?
MacEachren and colleagues (1999), based on the work of Michael Heim (1998), categorized extended reality technologies by “the four I’s” – immersion, interactivity, information intensity, and intelligence of objects. It may be helpful to consider these factors as we discuss extended reality technologies. The most common way to categorize these tools, however, is via a continuum of immersion, from augmented reality (i.e., the overlay of objects onto the real world), to virtual reality (i.e., full immersion in an imagined space).
As it is the most commonly-used term and a sufficient label for this technology, we will use the term virtual reality when discussing immersive computer-modeled environments. Note, however, that some scholars have used the term virtual environments instead. A virtual environment is a defined three-dimensional computer-simulated environment that enables user navigation and interaction (Slocum et al. 2009). The reason this term is occasionally preferred over virtual reality is that virtual environments often depict imagined things—for example, by visualizing Earth’s ozone layer, which is not actually visible to the human eye, and thus not a part of reality (Slocum et al. 2009).
Applications of extended reality in geography include creating virtual cities, virtual field trips, digital globes, and more. Shown in Figure 8.8.2 is a Computer-Assisted Virtual Environment (CAVE) from the Idaho National Laboratory.
Similar projects have been developed here at Penn State. In broadest terms, virtual reality is being used to permit travel to locations otherwise inaccessible to users. Immersive Technologies for Archaeology [239], for example, is a project that brings users to the Mayan ruins at Cahal Pech in Belize. Future plans for the project include completing a historical model of the Mayan city—permitting users to view not only a place but also a time—that they would otherwise be unable to inhabit. Other projects, such as Visualizing Forest Futures [240] (VIFF; Figure 8.8.3) extend in the opposite direction, giving users a view of the projected future of forests, based on possible future climate scenarios.
Think back to the first topic introduced in this lesson: cartographic generalization. How do we resolve the differences between the premise of generalization and the goals of VR?
As mentioned previously, not all extended reality is fully immersive—in fact, the fastest-growing type of extended reality is augmented reality (AR). Social media applications such as Snapchat use augmented reality as entertainment value, but AR can also be used in education and research applications. Figure 8.8.4 below shows the application Obelisk AR, developed at Penn State. Users can use the app to interact with a real-world object (The Obelisk) in a mobile environment. Tapping on a stone on the Obelisk, for example, brings up a pop-up on the user’s smartphone screen that explains the type and origin of the type of stone selected.
Augmented reality has also shown incredible potential for navigational and wayfinding applications. Earlier in this lesson, we discussed personal navigational devices in the context of passive interactivity. In some cases, augmented reality has taken such navigational devices to the next level. Pilots, for example, are often assisted via heads-up displays (Figure 8.8.5). These displays show crucial information overlaid across the environment, providing a better decision-making tool than a separate digital display.
Heads-up displays are also used for automobile navigation; such displays are offered by some luxury vehicle navigational systems, such as the one in Figure 8.8.6. Unlike AR pilot navigation systems, automobile heads-up navigation displays are not in widespread use. Increasing accessibility and affordability of these technologies, however, may result in them being the way of the future.
Watch this video of a car using a heads-up navigation display as it drives through intersections and accelerates onto a highway (2:02).
In Lesson 8, we talked a lot about interactive maps, and how the recent proliferation of such interactive maps has brought the challenges of map generalization back into focus. To create an effective interactive map, cartographers must consider not only how a map looks at one scale and extent (i.e., as in a typical static map) but at all locations and every scale.
As suggested above, creating an interactive web basemap can be a challenging task. Fortunately, tools exist to make this process easier and more efficient. In Lab 8, rather than using ArcGIS Pro, we will be working in Mapbox Studio [248]. Mapbox Studio is an online mapping platform for creating custom interactive maps. Mapbox maps can be used on their own, but they are also often used as basemaps in web-mapping applications and interactive thematic maps.
Before getting into the details of working in Mapbox, please be aware that the software is different than ArcPro. You need to remove much if not all of how you approach working with data in Mapbox. Aside from layers, there is little else in common between the way that Mapbox and ArcPro handles data. Thus, expect that there will be a bit of a learning curve with this lesson.
A rubric is posted for your review.
Further instructions are available in Lesson 8 Lab Visual Guide.
Unlike our other labs, we will not be using ArcGIS Pro for this lab, so there is no starting file! Instead, we will design an interactive basemap with Mapbox Studio.
Mapbox is an interactive web map design platform; there are many examples on their site of the possibilities designing with Mapbox provides. Check out the Mapbox Gallery [20]!
As you likely notice in the gallery link above, some of the most visually appealing Mapbox maps are not what we would consider traditional basemaps. They take significant creative license with their color and pattern design, while still incorporating proper cartographic generalization and providing legible map symbols. Our goal is to do the same in Lab 8 - you will be creating a new interactive basemap inspired by a favorite piece of art/media/design - have fun with it!
The first step in this lab is to create a Mapbox Studio account. Use your university email account, which will enable all necessary features for free. Once you're logged in, you should see the starting dashboard (Figure 8.0). Once the start-up screen is visible, you will click in the "Start by desinging a map >" area.
In the next window to appear (Figure 8.1), select the New style button to create a blank template that you will use to build your own map.
With Mapbox, you work with styles. These are templates that get you started with your design. Note, that styles are easily changed. Having selected New style on the previous window (Figure 8.1), you will now use an empty template from which to build your own map. A new window appears (Figure 8.2) listing several defualt style templates. Don't start with any of the pre-made templates that are listed. Instaed, scroll down the list of available styles and choose the Blank style. Then, select the Customize Blank button.
After selecting the Cutsomize Blank map template option, the studio editor will load: you should see a empty space to fill. We will be adding and styling layers one-by-one to create our basemap starting with the background layer (Figure 8.3). Before continuing, note that the process we will use to build a map in Mapbox is different from when we designed basemaps in ArcGIS Pro, keep in mind that many of the same principles apply (e.g., arranging the order of layers to match the visual order of the layers the reader sees).
Start by designing the background for your map. You can alter your background's color, opacity, etc. To turn on the background layer for you to edit, select...
- Layers
- the "+" icon
- Convert to background layer option unser Source.
Figure 8.3 shows the default background layer style as a black background. You can and should change this default setting to something that is inline with your own design inspriation. For example, in Figure 8.4, I've selected a cyan or sea-green hue (click somewhere inside the color palette to set the color) as my background color. To change the background color, click on the square color chip next to the background layer. Note that you can also change the name of the background layer to something else by clicking on the word "background" at the top-left corner of the background window. Once you are finished with editing the background colot, exit out of the editing session by clicking on the "X" next to the Style and Select text in the upper-right corner of the background window. As the goal of this lab is to draw inspiration from a piece of art/media for your design, you will probably come back to alter the background layer more than once. I strongly encourage you to experiment with the many style options available first, and then go back later to make changes to your design at a later time.
Visual Guide Figure 8.4. Editing the map background layer.
One by one, we will now add additional layers to our map. Click the "+" icon. Note that the New Layer window that appears is listed under the Select data item. Under Source, select the None selected option. In the panel that appears, use the vertical slider bar to navigate down to the Mapbox Streets V8 source as shown in Figure 8.5. The styles of the various "street" elements are shown. We'll start by adding the admin layer (you may have to scroll down a bit to see the admin layer. Select the layer to load it to your layer list. This page [252] provides an overview of all the individual types of layers and their attributes that are available in the Streets V8 file.
Now that you have some data loaded, we can Style the new layer. Similar to how we edited the map background, we can edit the color, opacity, line width, etc., of the administrative boundary lines in this layer. Click on any of the options shown in the admin window and alter them as needed. Figure 8.6 shows the backtground color overprinted by darker blue county and state outlines. Note that as the admin layer only consists of lines, we cannot fill in the administrative boundaries - the fill for the administrative boundaries will be the color of your background layer. Of course, you can return and change the color of the background layer if you so choose. Next, we will change the color of the ocean by adding a water layer later.
Add another layer: this time, select place-label from the Mapbox Streets V8 data set. Click on Type to set the type characteristics for this layer. To begin, choose Symbol (Figure 8.7). This action will result in your place labels displaying as text rather than as another kind of symbol such as a circle.
To display the labels and edit their appearance, you need to edit the Style of the place-label layer. To do this, click on the black rectangle to the left of the T place-label in the layer listing. On the menu items that appear, notice that the Text field option. Click on this option to select a "Text Field" that will be used to supply the text for your labels. Think of this process as choosing a field header of an attribute table. Select the Insert a data field that contains the labels you want.
If you look down the list of available data fields, you start your labeling by looking at either the "name" or "name_en" data field (Figure 8.8), depending on whether you would like the place names to display in the local language or always in English, respectively. You should explore other data fields for additional labels.
Once you have selected a data field to label, you will want to Style those labels using the Text, Icon, Position, and Placement options on the T Place-label window. Remember to work with the zoom across range option to maximize the visilibilty of your labels at different zoom ranges.
Once you've made a bit of progress on your map, you should Publish it. To do so, click on the Publish button in the upper right-hand corner of the screen. The Publish Blank window appears (Figure 8.9) - this also saves the map in your account. You should re-publish as you work, similar to how you regularly save your work while using desktop software.
Once your map is published to the web, it will be available via a shared URL so that anyone with the link can view your map. Set your style to public. To Share your Mapbox creation through a URL, click on the Share button to the left of the Publish button (Figure 8.10a). On this window, you can Make public your Style URL. This is the URL that uniquely is attached to your Mapbox style. Note in FIgure 8.11 the Allow copying (at the top center of the window in Figure 8.10a) is how you can copy the URL, submit that URL with your deliverable, and share yoru Mapbox creation with others in the class for their peer review efforts.
Once you have Shared your style, you can let anyone with the link can view your map (Figure 8.10b). When you Allow copying with others in the class, make sure to copy the URL link that starts with https://api.mapbox.com/styles/v1/...
In addition to adding and styling new layers, Mapbox permits a lot of creative customization. Though it's not required, you may want to add a creative font to your map - such as one that is distinctly used in your inspiration source. Many open source fonts are available: here I've uploaded an OTF (OpenType Font) file, which I downloaded online from fontspace [253]. This font is called the priest. The font looks kind of strange but it could work with a sci-fi kind of theme. To load the font for a particular layer, click on the Font option in the Style item (Figure 8.11). Select the "Manage font in your account: link at the bottom of the listing of available fonts. Note that TTF (TrueType Font) files will work as well - try Google Fonts [254] for free downloadable font files.
As shown in Figure 8.12, this custom font creates an interesting look and feel to the map.
Once you've made some progress on designing your basic map style, you should make some more detailed edits. Remember that we are creating a multiscale map. You may, for example, want your place labels to appear at a different point size based on the map's zoom level.
In Mapbox, certain layers have a minimum zoom level. For example, administrative boundaries are seen at all zoom levels (so, this layer has a minimum zoom level of 0), poi_labels have a minimum zoom level of 5, and large buildings have a minimum zoom level of 13 (all buildings show up at zoom level of 16).
To style place labels, use the "Style across zoom range" function (Figure 8.13).
Style across zoom range permits you to alter the look and feel of a map symbol dynamically as the user zooms in and out of your map. Figure 8.13 shows that you can choose from different rates of change (e.g., linear - where symbols change gradually, or step - where symbols change abruptly at zoom levels you define). You don't have to do anything overly complicated - the goal is just to map your map look nice at small, medium, and large scales.
Tip! Look at other online interactive basemaps (e.g., Google Maps), as well as the Mapbox examples listed at the top of this guide for ideas about what symbols should appear at which scales, and how they might look best.
In addition to styling your layers based on the zoom layers, there is another type of condition you should utilize: the type of feature. Some layers contain contain multiple types of features - road is a prime example. Mapbox permits you to style features differently based on a data condition. In this example (Figure 8.15) roads with the class "primary" are assigned to be sized larger (2 pt.) than those with the class "secondary" roads (0.75 pt.). This weight difference is very similar to when we used TNMFRC codes (e.g., " 1 = Interstate," " 4 = Local Road").in ArcGIS Pro to assign different road types to label classes, and sized the labels based on their classification.
It is important to note that the naming convention of the different types of roads in Mapbox does not necessarily follow what you are accustomed to with interstate, secondary roads, and so forth. You should explore the different naming conventions to make sure you understand to what each class refers. This page [252]may be of some use to you as you explore the different classes of roads.
Once you have added roads, buildings, etc. you will want to add additional labels to your map. We previously added the place-labels layer to our map, which is a layer intended just for labeling. However, we can also add labels using other layers. To label roads, for example, add another instance of the road layer using the "+" button, but this time select Symbol as the Type instead of Line (the default). See Figure 8.16.
Most of your layers will come from the Mapbox Streets source, but you should also add one terrain layer. Just as in Lab 6 (terrain visualization) you should use transparency and layer ordering to establish your terrain visualization effective. As usual, you want the terrain layer to be visible, but not visually overwhelming. You can drag and re-arrange layers in the left pane of Mapbox studio just as we did in ArcGIS Pro. Note that the design shown in Figure 8.17 is not that good. You can do better!!!
As this is likely to be your first time using Mapbox Studio and the interface is different, take your time to become accustomed with how Mapbox works. For more practice, you might want to run-through an online tutorial such as Mapbox's Create a Custom Style Tutorial [255]. Though this will be informative and helpful, you should still follow this guide and start from blank, rather than using a pre-made Mapbox template or the color suggested in an online tutorial.
Here is a video tutorial [256] on how to filter what appears on your Mapbox creation.
Here are a series of video tutorials [257] on some basic manipulations in Mapbox by a fellow colleague of mine: Ian Muehlenhaus.
Map design is an iterative process and it may take time for you to get a design you are happy with - be patient with yourself and remember to draw ideas from other maps, your media/art inspiration, and course content.
Credit for all screenshots is to Fritz Kessler, Penn State University. Screenshots and data from Mapbox Studio.
You've reached the end of Lesson 8! This lesson, we discussed the related topics of cartographic generalization and multiscale maps, as well as how these concepts are integral to creating effective interactive web-maps. While introducing new mapping techniques (e.g., animated maps, virtual reality) we discussed both the opportunities and challenges that new technologies in map-making provide. Using Mark Monmonier's conceptualization of fast maps, we discussed how even static maps have taken new forms in recent years due to the ability of social media to spread such maps fast, far, and wide.
In Lab 8, we used a new cartographic tool—Mapbox Studio—to create an interactive basemap inspired by a favorite piece of art. In Lesson 9, we continue along this trajectory of focus on interactivity and web-based map dissemination. We move next from creating an interactive basemap to an interactive thematic map and visual graphics with data visualization software Tableau.
You have reached the end of Lesson 8! Double-check the to-do list on the Lesson 8 Overview page [258] to make sure you have completed all of the activities listed there before you begin Lesson 9.
Welcome to Lesson 9! Last week, we discussed some of the new technologies that have been influential on current trends in cartography, including interactive and animated maps, 3D visualization, and augmented reality. While interactive and dynamic maps present a myriad of opportunities for creating new and exciting designs, they also introduce many new challenges. Studies of interactive maps draw from research not only in cartography and psychology but in other cognate fields such as human computer interaction (HCI), human factors, and usability engineering. We will discuss various approaches for studying dynamic maps in this lesson.
Unlike the static maps that we have focused on for most of this course, maps we discuss here are dynamic—they change based on interactions (either active or passive) by the map reader. In such cases, we begin to consider the map reader instead as a map user. Additionally, as these maps typically appear alongside other media (e.g., supplemental charts, article text, videos), we consider these map-adjacent elements and how they influence the user experience. In Lab 9, we put this knowledge to use and design an interactive data visualization story with the visual analytics platform Tableau.
Action |
Assignment | Directions |
---|---|---|
To Read |
There are no external required readings for Lesson 9. Instead, you should explore in-depth the links included in this week's lesson content. In particular, please explore the three links to graphic compilations (NYT [259]; Washington Post [260]; Nat Geo [261]) and the Tableau Stories about AirBnb in Portland [262] in the Data Journalism section. Additional (recommended) readings are clearly noted throughout the lesson and can be pursued as your time and interest allow. |
|
To Do |
|
|
If you have questions, please feel free to post them to Lesson 9 Discussion Forum. While you are there, feel free to post your own responses if you, too, are able to help a classmate.
We often consider how our map readers might interpret or respond to a map we make. Predicting and designing our maps to account for this is a complex problem that we have discussed throughout this course. When making maps, we often must choose a suitable projection, symbolize data appropriately, visualize additional elements such as terrain, and so on. We also account for contextual factors: for example, we might expect our map readers to be stressed or working within time constraints. We may also need to design for media-based constraints such as black-and-white newspaper printing, or for challenging viewing scenarios, such as small sizes (e.g., in an academic journal article) or far distances (e.g., in a slideshow presentation).
You might recall the maps in Figure 9.1.1 from Lesson 1 - each was designed with a different type of map reader in mind.
Figure 9.1.1 shows how minor alterations to a static map (here, technically still-frames of a larger map) can make it more suitable for a given map audience or purpose. Last lesson, we introduced interactive maps—maps that change based on some form of user input. This new realm of interactive mapping has turned our focus from the map reader to the map user (Roth et al. 2017). We now must consider not just how our map’s audience will interpret the map we design in a single state, but how they will interpret it as they use it, which is to say, as they alter it. An interactive map must work not only in one state, but ideally at every state that might be reasonably encountered by the map user. This is no small task.
Even basic interactions such as panning around a slippy map can introduce challenges. Figure 9.1.2, for example, shows two locations on an OpenStreetMap basemap, both at a 1:5,000 scale.
These maps are shown at the same scale but appear vastly different—and this makes sense, given that they are very different places. What this example highlights, however, is the variety between locations that pan-able maps must often be designed to cover. Web maps typically cannot be designed separately for each location (imagine the time that would take!) so cartographers use generalization algorithms and design rules to ensure that their maps will work at locations rural and urban, near and far, and at scales both small and large.
Panning (i.e., moving the map to another location) is among the most basic functions performed with interactive maps. Additional functions such as filtering and route-planning introduce further complexities to interactive map design. For insight on how to best support such tasks, cartographers have turned to the study of usability. Usability is defined by the International Organization for Standardization (ISO 9241-11:2018) as “the extent to which a system, product or service can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use.” Designers of websites, mobile apps, and many other technologies consider system usability when building their products. Though it is a subject with a rich history and many facets, Jakob Nielsen’s (1994) usability heuristics [263] provide an excellent foundation for assessing the usability of a system (such as an interactive map).
Figure 9.1.3 demonstrates two of Nielsen's usability heuristics: error prevention, and consistency and standards.
View Nielsen’s usability heuristics [264] online. Open ArcGIS Pro, and search for examples of these heuristics in the interface. You might also try this out with another favorite (or least favorite) software program. Which heuristics are implemented? Which are forgotten?
As suggested by the ISO (2018) definition, an important component of usability, and one that ought to be considered when implementing the usability heuristics is the idea of context of use. For example, a routing app might be designed specifically so that the interface can be safely manipulated (or not) while the user’s vehicle is in motion.
Despite the importance of context in designing usable systems, a significant amount of scientific research related to usability has focuses on developing more generalizable findings, such as whether users can identify changes in animated maps (Fish, Goldsberry, and Battersby 2011). When we consider how to assess maps in terms of their usefulness, it is helpful to distinguish between these two primary approaches: traditional, experimental research intended to elucidate generalizable insights, and design studies that focus on context-specific design. Roth and Harrower (2008) describe these sorts of studies as a continuum from controlled experimentation to usability testing. Despite the helpfulness of conceptualizing cartographic evaluation methods as existing along a continuum, we discuss these methods as falling more generally into one of two categories (1) experimental studies, and (2) design studies, for simplicity and brevity.
As noted in the previous section, experimental studies seek to identify generalizable findings. These studies draw heavily from work in psychology, a discipline with a rich history of closely-controlled experimental research. Research conducted by Fish, Goldsberry, and Battersby (2011) on change blindness is a helpful example of experimental research in cartography.
Consider the maps in Figure 9.2.1 below – after viewing these animated frames, do you think you would remember which states changed from the first (left) to the second (right) frame?
Fish, Goldsberry, and Battersby (2011) found that not only did their participants often incorrectly identify which locations had changed from previous animation frame, they were consistently overconfident in their reports. A suggestion made by the authors to mitigate this effect was to incorporate tweening, or gradual graphic transitioning between animation frames, into animated map designs. This suggestion is applicable to a wide variety of animated mapping contexts.
Similar studies have been conducted on other aspects of map design. Limpisathian (2017), for example, studied the influence of visual line and color contrast on map reader perceptions of feature hierarchy and aesthetic quality. Unlike Fish et al., who conducted their research with participants in a lab, Limpisathian recruited and tested participants using the crowd-sourcing platform Amazon Mechanical Turk (MTurk). Such platforms have become increasing popular in recent years as—despite their shortcomings—they enable researchers to run large studies with more diverse sets of participants and at lower costs.
Experimental studies often use web surveys, which can measure task (e.g., map data retrieval) accuracy and speed. Some surveys take advantage of new technologies such as eye-tracking, which measures fixations of the human eye. Griffin and Robinson (2015), for example, used eye-tracking to measure the efficiency of color and leader-line approaches when highlighting geovisualizations. Eye-tracking is a popular method for understanding user response to design, and is regularly used by web design practitioners and in marketing research. Figure 9.2.2 shows an example record of eye-tracking from a study performed on the Healthcare.gov website. Similar studies have been conducted with maps and other spatial tools.
Fish, Carolyn, Kirk P. Goldsberry, and Sarah Battersby. 2011. “Change Blindness in Animated Choropleth Maps: An Empirical Study.” Cartography and Geographic Information Science 38 (4): 350–362. doi:10.1559/15230406384350.
While experimental studies focus on producing generalizable findings (e.g., “people suffer from change blindness when viewing animated maps”), design studies focus on more specific use contexts. The goal of these studies is generally to improve a specific map or mapping product. Testing often begins early in the design stage, with preliminary design sketches and/or paper prototypes (Figure 9.3.1). Paper prototypes are generally preferred to more formalized mock-ups early in the design process, as they cost little to create, leaving designers more willing to change their designs in accordance with suggestions by testers. Research has also found that testers of "sketchy" designs and paper prototypes are more likely to elicit big picture design suggestions than more formalized prototypes (Dykes and Lloyd 2011). This is because test users are more able to focus on the overall functions of a tool when they view it as unfinished—they are not distracted by small design details (Dykes and Lloyd 2011).
As design studies focus on a specific use context, it is important to test with target users (i.e., the intended users of the product) whenever possible. A map designed to be used by police officers, for example, will likely require input from these users to be made sufficiently useful in that context. A popular mantra in usability research is this: you are not your users. When designing a map intended for use by the general public (e.g., Figure 9.3.2), it might be enough to test your design with a group of college undergrads for course credit, or through a crowdsourcing platform such as Amazon Mechanical Turk. For more specialized contexts, recruiting those target users is necessary.
Roth, Ross, and MacEachren (2015) emphasize this importance of involving target users throughout the map design process. In their work designing an interactive mapping tool to support the needs of the Harrisburg, PA Bureau of Police, they suggest an iterative approach to system design. They outline three U’s of interactive map design: user (i.e., considerations of target users and use-case scenarios), utility (i.e., whether the map performs the tasks that its users require), and usability (i.e., whether the tool’s functions align with usability design principles/heuristics).
When we talk about interactivity in maps, we must consider not just user interactivity within maps, but interactively among maps, as well as with other tools and visual graphics. Interactive mapping has played an important role in the field of visual analytics, defined as “the science of analytical reasoning facilitated by interactive visual interfaces” (Thomas and Cook 2005).
Recall the Cartography Cube from Lesson 1 (review this concept in the Communicating with Maps [268] section). Most of the maps we have designed thus far would be considered to be in the communication (public, static, and intended to present known information) corner of the cube. Visual analytic tools typically belong in the opposite corner—these tools are characterized by high human-map interaction and are often designed with private data or data that is otherwise meant for domain experts. They also focus on revealing unknowns (i.e., generating insights), rather than communicating known trends.
One domain in which visual analytics has been particularly popular is in public health and epidemiology. An example tool is shown below (Figure 9.4.2). The Pennsylvania Cancer Atlas is an interactive tool designed at the GeoVISTA Center at Penn State, with assistance from the Centers for Disease Control (CDC) (Bhowmick et al. 2008). The atlas includes a choropleth county-level map of Pennsylvania, coordinated charts and tables, and filtering and selection options to compare data across the views. In the view shown below, for example, Bedford County has been selected on the map by the user, and the scatterplot and table have been highlighted to focus on that county as well. This connecting of multiple visual depictions of data is called coordinated views.
A more recent example is FluView, a visual analytic dashboard designed by the CDC for analyzing data related to incidence of the flu in the United States. FluView is shown in Figure 9.4.3 below—you can try it out by selecting the link here: FluView [270].
A demo of a more complex geovisualization built around visual storytelling, Detecting Disease Spread from Microblogs, is shown in the video in Figure 9.4.4. below:
Selecting ‘lil’ microblogs (0:02)
The first stage of our analysis involved identifying the key words and phrases that we thought were associated with the epidemic, this allowed us to select only those blog entries that we thought were relevant for the analysis of the disease.
Where, when, what (0:13)
Our main application comprised three views of the blog posts, firstly one showing where they occurred, secondly one showing when they occurred, including the associated weather over this timeline, and thirdly, the posts themselves. The distribution of posts shown on the map indicate a concentration around the hospitals, this led us to believe that at least some of these posts were second or third entries from people who’d already fallen ill elsewhere. We could confirm this by examining the history of the people who tweeted on the map. Here we see all posts by the same poster, indicating that they’d tweeted several times about the same illness. This led us to filter our data so that only the first entry from each poster was shown on the graph, here shown by red bars, and on the map we see that there are no longer any concentrations around the main hospitals, indicating that people first posted when they became ill, away from the hospitals.
Ground zero (1:21)
The timeline shows very clearly when the epidemic first starts, about the 18th of May. We can do a temporal selection on the data to find out how to disease begins to spread from that point. The timeline shows data grouped into bins of 6 hours. To identify ground zero we can change the resolution of the bins to a much finer-grained analysis. By performing a temporal selection at this new resolution, we begin to see what happens at the start of the outbreak. Looking at the map view as we move through time, we begin to see the first outbreaks of the disease in the downtown area. This led us to believe that there were three areas in the downtown region where the disease first emerged: The Vastopolos Dome, next to this Vastopolos Hospital, and around the Convention Center. We also see some spread towards the riverside of the Dome.
Spread and containment (2:18)
To be sure that we were viewing the real spread of the disease, rather than the propensity to microblog, we created a chi-expectations surface of the region, where dark green areas show a greater than expected density of ill posts, and purple areas show a less than expected density. In addition to the Dome, the Hospital, and the Convention Center, this also reveals that Eastside has a greater than expected density of incidences. The third region to show the spread of the disease is toward the west of the region on the banks of the river. This is in contrast to the southern areas of downtown and uptown area, which seem relatively unaffected by the disease. Finally, we summarize the distribution of points using a standard ellipse. This allows us to examine how the disease spreads over time, by performing a temporal selection on the bar chart at the bottom, and then moving through time, we can see how that standard ellipse, which gets dark green with a high concentration of the disease, is dragged towards the southwest by instances of a completely different disease, associated with the river. By filtering posts that show sickness, diarrhea, and stomach cramps, we clearly see the river association of the disease, which started at 2am on the 19th. To examine whether there’s any spread beyond the length of the river, we can perform a spatial selection of just those points associated with the river, and examine how that changes over time. Doing so reveals that while there’s a high concentration towards the northeast of the river, this doesn’t move downstream over time. We can therefore be confident that the disease is relatively well-contained.
Though health and public safety applications are popular uses for (geo)visual analytic tools, they have been used in many domains. Figure 9.4.5 below shows the geovisualization tool MapSieve, designed for analyzing spatial patterns of student engagement in online courses taken by students all over the world.
While the tools above focus on fairly complex data that often require domain knowledge for effective interpretation, similar visualization tools are also often used in more fun, less serious contexts. Figure 9.4.6, for example, shows a Tableau (data visualization software) dashboard that visualizes AirBnb data in Portland, Oregon. We will take a closer look at dashboards like this later in this lesson.
Similar interactive tools are often designed for mapping election results or other data of public interest. Graphics are often accompanied by a significant amount of text, both within the main view as explanatory text, or adjacent, to tell a story supported by the data. We discuss this more in the next section: Data Journalism.
Bhowmick, Tanuka, Anthony C Robinson, Adrienne Gruver, Alan M MacEachren, and Eugene J Lengerich. 2008. “Distributed Usability Evaluation of the Pennsylvania Cancer Atlas.” International Journal of Health Geographics, no. February 2015. doi:10.1186/1476-072X-7-36.
As demonstrated by the Portland AirBnb example, interactive maps designed for public consumption often do not stand alone. Except in the case of very simple data visualizations, these maps and graphics tend to be accompanied by additional text, both within the visualization interface and outside. Such maps are often included—in either static or interactive form—in the type of articles and other media described as data journalism.
Data journalism is a general term that refers to the increasing influence of numerical data in news reporting; data are often reported and/or visualized alongside articles and reports disseminated to the public. Though data journalism does not necessarily have to include visual depictions of data, it often does—and for good reason. Visual graphics tend to captivate readers, and charts, maps, and graphics are much better at communicating data at a glance than data tables and spreadsheets alone. The article in Figure 9.5.1 is an example of this. The article includes a large map, as well as a set of small multiple maps, to visualize the geographic distribution of ammonia. Article text gives the reader additional information about the ammonia gas.
Journal outlets such as the NY Times, Washington Post, and National Geographic are among those creating high-quality graphics and accompanying article content. Visit the links below to see examples:
As demonstrated by the links above, media outlets frequently report on important and emotionally-engaging information. Journalists take on the challenging job of reporting this information to the public. Often, pairing interactive maps and graphics with carefully curated text is the most effective way to do so.
Think back to MacEachren’s Cartography Cube. Where would you place the maps/graphics included in the articles referenced in the links above?
Given recent trends, including as the proliferation of interactive maps and visual analytics, cartographers have begun to focus more on maintaining a balance of text, graphics, and other elements in their work. Think back to our discussion in Lesson 2 of frame lines and neat lines for map layouts—such simple guidelines seem almost irrelevant in the context of data journalism and interactive map making. While cartographers still face traditional design constraints when creating maps for print (e.g., magazine spreads, print advertisements), they must now also work with more complex design formats such as infinite scrolling web pages and interactive dashboards.
In previous lessons, we discussed the importance of design thinking that reaches beyond the map—configuring page layouts and explanatory text in a neat, usable, aesthetically-pleasing fashion. Given our current focus on the map user, note that ideally this design thinking ought to be implemented at all stages of map interaction. For example, see Figure 9.5.2. Shown in this view is the map “on-hover,” which means that the user has hovered their cursor over the point that is highlighted. The tooltip which appears (Figure 9.5.3) must present an amount of information that is adequate but not overwhelming for map users. It could be argued that this is not successfully accomplished here—the coordinate location is likely unnecessary information, and the addition of a short description of the property would assist the map user.
The visual information-seeking mantra, first proposed by computer scientist Ben Shneiderman, is frequently referenced by information designers: “Overview first, zoom and filter, and details-on-demand” (Shneiderman 1996).
We will use the Portland Airbnb Tableau dashboard to explore this mantra in practice. First, in Figure 9.5.4, the starting view of the dashboard, which shows all of the listings in Portland: overview first.
From the starting view, the user can zoom in and/or pan around the map, and filter the map data by selecting a category of interest. The tool state in Figure 9.5.5 shows the view after the user has zoomed into the map, and selected the "private room," room type. This data could be further filtered by by selecting a property type, such as "hotel-like property." This is the second stage of the mantra: zoom and filter.
Figure 9.5.6 shows the view in 9.5.5 upon mouse hover of this hotel-like property near the river. ID numbers for the host and listing, as well as lat/long coordinates, are given in the tooltip. This is the final element of Shneiderman's mantra: details-on-demand.
Play around with this Tableau Story, Airbnb Data in Portland [277]—in addition to helping you understand the concepts presented in this lesson, it may give you ideas as you work on Lab 9.
Small snippets of text such as tooltips, titles, weblinks, and error messages associated with your maps will often be designed by you, the cartographer. Such text is often called microcontent, and despite its minimal nature can have a large impact on user interpretation of your visualizations.
The Nielsen Norman Group provides a helpful reference on how to write such content here: Microcontent: A Few Small Words Have a Mega Impact on Business. [278] Their site is also an informative reference for many aspects of usability and user experience design.
Shneiderman, B. 1996. “The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations.” Proceedings 1996 IEEE Symposium on Visual Languages, 336–343. doi:10.1109/VL.1996.545307.
The visual analytic tools we have have explored thus far include both maps and graphs, and these different data visualization elements have been connected via coordinated views, permitting user filtering, zooming, and more. Given the limited space available in these dashboards—particularly if they are intended for viewing on small, mobile screens—an important question surfaces: do I need a map at all?
When designing data visualizations, maps often provide an invaluable source for insight generation. However, they are not necessarily always the best choice for your data—even if the data contain spatial information. View the dashboard below in Figure 9.6.1.
This dashboard does not contain a map—and though it’s possible that adding one might provide additional information or interest, its current form gets across the core message: drug overdoses are increasing in Philadelphia, and this is being driven by opioids in general, and fentanyl in particular.
Given the increasing ubiquity of GPS and other location-based technologies, data that contains a geographic component (e.g., state, county, coordinates) is fairly easy to come by. Still, this does not mean that creating a map is always the answer. Imagine, for example, if you had collected data on the rate of Alzheimer's disease by state. Were you to map this data, popular retirement states such as Florida and Arizona would likely jump out—not because there is anything inherently unhealthy about those locations, but because their populations are significantly older. To eliminate the effect of this confounding variable, you could map age-adjusted Alzheimer's rates instead. It's important to consider, however, whether this would be the most informative way to visualize your data. If you were you simply hoping to educate people about Alzheimer's, a graph or chart comparing Alzheimer's rates by gender, age, race, or socioeconomic status might serve your purposes just as well.
Conversely, there are many data visualizations that unfortunately treat space (i.e., geography) as just another variable. For an example, view the dashboard in Figure 9.6.2 below. The designer of this dashboard chose to visualize states as list of values, rather than to create a map. Though this is not inherently incorrect, it is a missed opportunity to provide the user with an at-a-glance understanding of spatial patterns in occupational data. Sure, the user could still pick out individual data values, or compare average annual earnings state-to-state. But data visualization (cartography included) is about making complex data clear—if your visualization is no more useful than its source data table, then why design it at all?
Your final lab assignment in this course is to design an interactive geovisualization using Tableau. While this lab draws heavily on concepts discussed in Lesson 9, you will be incorporating knowledge from throughout the course in your design. Unlike other labs, Lab 9 is a two-week assignment.
In Week One, you should develop an idea and gather data for your lab, and complete the example Tableau Story tutorial (Visual Guide Part 1). This tutorial is ungraded, but will teach you the basics of working in Tableau. You will then create your own Tableau Story using your own data. The Lab 9 Visual Guide Part 2 contains tips and tricks for working in Tableau beyond what is covered in Part 1, and a wealth of additional resources are available via the web.
Submit the link to your Story (hosted on Tableau Public) as a text comment. There is no PDF deliverable for this lab.
A rubric is posted for your review.
Please refer to the Lesson 9 Lab Visual Guides: Part 1 and Part 2.
In this lab, we will design an interactive geovisualization with Tableau. The final result will be a Tableau Story similar to the one about Airbnb data in Portland we discussed in Lesson 9.
To begin, open the Age_andSex_AFF_ACS_2017 [286] Excel file. This file has multiple fields (columns) of data for each state in the United States. It was created by making minor edits to a CSV file downloaded from the US Census Bureau's American Community Survey (ACS). If you're not sure what data to use for your own project, the ACS is a good place to start.
The most important component of this Excel sheet is the State column—Tableau will automatically recognize and map several geographies, such as States, Countries, Zipcodes, and Coordinates (lat/long). You may choose to map another geography (e.g., counties, census tracts, block groups) for your own Lab, but using one of these geographies is more advanced and will not be covered here.
Open Tableau Desktop and Connect to the Excel File (Figure 9.2.). The file has already been formatted properly for import. If you like, you can select "Extract" to extract the data. If not, you will be prompted to do so later, before publishing your Story online.
Select Sheet 1 at the bottom of the page to open a Tableau worksheet.
Before continuing, save your file as a Tableau Workbook file (the default file type). As with projects in ArcGIS Pro, you should regularly save as your work.
You should now see a screen similar to the one in Figure 9.3. State should be listed among your tables, and your measures should list the many fields of data that were included in your Excel file.
The distinction between table and measures in Tableau is important. A table is an element, such as a state, year, company, etc. that you are interested in viewing data about. A measure is that data, such as % insured, or a number of products sold. For geographic data, a table is always the geographic unit (e.g., state, country) and a measure is the data to be mapped. Tableau often correctly identifies the table and measures in your data, though occasionally you may have to convert one to the other. In this example, all measures and tables were correctly identified.
Now it's time to make your first map! Click and drag the State table into the middle of your worksheet (Figure 9.4). Tableau will automatically recognize this geography and create a map.
Drag a measure of interest onto an appropriate visual variable in the Marks section (Figure 9.5). In this example, “Percent Female; Estimate; AGE 85 years and over” is dragged to the Color box. You might choose another mark (symbol) type such as size if you were mapping count values (such as the TOTAL number of 85+ yr old females, rather than the rate).
Select “Edit Colors” to choose a different color palette – remember to choose a color progression appropriate for the progression of your data! You can use the “advanced” menu to make further edits.
Recall that the focus of this lab is to create an interactive map/dashboard with coordinated views. Here, we create a bar chart with the same data as our map. The intent of this is to show the same data in two different ways. Eventually, we will connect the map and graph so that the user can explore one via the other.
To create a graph, first, create a new worksheet. Then drag one measure (e.g., “Percent Female; Estimate; AGE 85 years and over”) and one table (here, "State") to Columns and Rows section. It doesn't matter which is which - switching them will simply change the orientation of your graph.
You may notice that when you add your measure (here, % female 85+) to a graph/chart/map in Tableau, the default measurement is SUM (see Figure 9.7). Since in our data we have only one value per state, the sum is equivalent to the original value. Thus, changing this is not necessary. If you had an Excel file with multiple rows for each state, Tableau would sum those values and display that calculated value - you may, in that case, want to display the average in each state instead. You can change how your measures are calculated by clicking on the colored green oval "pill" of the measure you want to change.
Once we've created a graph, it's time to add color! Drag the same data measure from the sidebar to the Color box in the Marks section to color your bars according to that data—as the length of the bar already represents this value, adding color here is called dual encoding. Edit your color scheme so that it matches the one from your map. Your color schemes (map and graph) should be equivalent, as we are only going to create one legend for our dashboard.
The last step before adding our visualizations to a dashboard is to clean up their design. Add chart/map titles (if you wish), shorten and clarify axis labels, and simplify tooltips (Figure 9.8). Adjust font size and style as appropriate.
Click the "New Dashboard" icon at the bottom of the screen (Figure 9.9) to create a Tableau Dashboard. This is what we will use to connect our map and graph. Drag your two worksheets onto the dashboard and re-arrange as you wish. Experiment with different arrangements of elements (e.g., map, graph, legend) on the dashboard.
Note: I have increased the size of my dashboard for demo purposes - you will likely want to use a standard size, and these are listed in the Size dropdown menu. The size of your dashboard on your screen vs. on the web will depend on the resolution of your laptop screen. It may take a bit of trial and error to get yours to appear the correct size.
You can use the Objects menu (bottom left) to add other elements, such as images, text, and blank layout elements to your dashboard. In Figure 9.10, a text object was used to add explanatory text, and an empty object was used to insert a margin above the legend.
To connect your map and graph, use the Actions menu (Figure 9.11). Here, Highlight is used to connect your map and graph upon user selection of an element on either. This is the default action if you do not customize anything, but you are welcome to use a different action or actions if you choose.
Your graph and map should now be connected! Example connection upon user interaction shown in Figure 9.12.
Once this is complete, create two more worksheets (one map; one graph) for another measure (I chose % Males 85+). When this is complete, you will have two similar dashboards that visualize your two chosen variables (e.g., % Females 85+, % Males 85+). You may choose to use the same legend and color scheme for both dashboards or switch it up – just keep in mind design principles from this course.
Select "New Story" at the bottom of the page to create a Tableau Story. While Tableau dashboards can contain multiple worksheets, Tableau stories can contain multiple dashboards.
Here, we create simple Story: Drag your first dashboard to the center of your new Story, then add a New Blank Story point. Use that new Story point to add your 2nd dashboard. Add an overall Story title and Story point titles (shown in the clickable grey boxes) as you wish.
Once you're happy with your Story design, you're ready to publish to Tableau Public. Make sure you've saved your work first! You will need to sign into (or create) your Tableau Public account before you can publish your work.
To publish to Tableau Public, use the menu structure to go to Server -> Tableau Public -> Save to Tableau Public (Figure 9.14). If you make changes to your Story, you can "re-publish" it at any time to update the online version.
When publishing your Story, you may be presented with a notice that a Data Extract is required (Figure 9.15). If so, simply select Create Data Extract in the window, and save the extracted data as suggested. Then repeat the above steps (Figure 9.14) to publish.
This example can be found here: Aging in the United States, a Tableau Story by Cary Anderson. [281] Use this link to check and see how your results from this visual guide/tutorial compare! If you do not see the "Create Data Extract" option as shown in Figure 9.15, then look at this link [287] to a help page with steps on how to do it. About midway down the webpage is the "Create an extract" section and it worked to publish part 1.
Credit for all screenshots is to Cary Anderson, Penn State University. Screenshots from Tableau Desktop, data source: US Census Bureau, the American Community Survey.
There are several geographies that Tableau will automatically recognize: States, Countries, Zipcodes, and Coordinates (lat/long). The process for Countries and Zipcodes is the same as States, which was demonstrated in the tutorial (Visual Guide Part 1). Shown below is a spreadsheet of school locations in Philadelphia, PA, which includes point location data (lat/long coordinates).
Lat/long data in the above spreadsheet is shown in the X and Y columns. To make things easier, I'm going to rename these latitude and longitude. Clean up anything else as needed. I edited one zipcode that was improperly formatted.
Recall from Part 1 that dimensions are elements such as States that you want to visualize data about, and measures are the variable data we want to display. In this example, Tableau categorized latitude and longitude as measures, but we would like to use them as dimensions. To change a data type from a measure to a dimension, simply drag it to the dimensions section.
Once you have longitude and latitude dimensions, you can add them to the columns and rows sections (respectively) to create your point locations map! Use the Marks menu (Figure 9.18) for adjusting symbol size, shape, etc.
We previously discussed cleaning up tooltips in Tableau. Tooltips will automatically describe data displayed on your map, but you may want to add additional data to them. For example, here we have mapped all schools and colored them by their school type. If we want to add the school name to the tooltip but not to the map, we can drag that measure directly to the Tooltip box in the Marks section to add it.
There are many options for altering the design of graphs and charts as well. For example, you can "drag to resize" a chart (Figure 9.21). Note in this example that the categorical color palette matches the one from the map - this is a good way to make comparisons easier for your viewers.
You can also change how your measures are calculated (and thus, displayed). In Part 1, each dimension (state) had only one row in the data table, so it didn't matter whether we mapped the sum or the average - they were the same value. In this second example, we could visualize the sum, or total enrollment for each type of school, (as in Figure 9.21) but it would be more interesting to show the average enrollment for each school type. You can alter how a measure is calculated as shown in Figure 9.22 below.
Though the examples in the lab instructions only use maps and bar charts, there are many other charts available in Tableau that you might create. The example shown below is a treemap. Click the "Show Me" menu to view all chart options. You also can look online for more advanced customization options - for example, there are tutorials that discuss how to create hexbin maps.
The data you use may have multiple types of geographies. For example, though we have only mapped the schools from this Philadelphia Schools dataset as points so far, the data also listed the zipcode for each school. Zipcodes are one of the geographies that Tableau automatically recognizes, so we can use this to create a graduated symbol or choropleth map to compare with our point locations map.
To do this, as we did with the States dimension in Part 1 of the Visual Guide, simply drag the Zipcode dimension onto the map.
To symbolize this map, drag your measure of interest to the color section of the Marks window, as shown below. Tableau will likely automatically created a choropleth (they call this a "filled map") but you can change this as well.
Tableau provides quite a few options for advanced customization. For example, you may want to use a color palette, such as one from ColorBrewer [288] or CARTOColors [289], that is not included in Tableau. Follow the instructions here: Create Custom Color Palettes [290] to add your own custom color palette.
Another (simpler) customization you can make is to change the basemap used in your Tableau map. The default light grey Tableau map, for example, does not provide much locational context. We can instead use a basemap from Mapbox. Log into your Mapbox Studio account, and find a map - you should select "Share & use" as shown in the top-right corner of Figure 9.27 below. This will lead to a URL you will paste into Tableau.
Instructions are also listed in Mapbox for adding a map to Tableau. You should follow these instructions (click Map > Background Maps > Map Services > add > Mapbox Maps; paste the share URL). Once you paste the URL, the other fields in Tableau will auto-populate.
Tableau dashboards can become fairly complex, but for this lab, we will not include too many pieces. Figure 9.29 is an example of a dashboard that compares two maps and a bar chart. See Part 1 of this visual guide for a refresher on how to create a Tableau dashboard.
As you work, try out different arrangements of maps and charts, as well as different chart types. In Figure 9.30 below, the graphics shown in this guide have been divided into two dashboards: one about enrollment by zipcode, and one about enrollment by school type. Each of these dashboards was then added to the overall Tableau Story as a story point (the two story points are Enrollment and Types).
Once you've designed your story, the last step is to shorten and clarify axis labels, simplify tooltips, adjust font sizes and styles, etc.
This example can be found here: Schooling in Philadelphia, a Tableau Story by Cary Anderson. [291] Note that you can download the workbook to practice with if you so choose.
Credit for all screenshots is to Cary Anderson, Penn State University. Screenshots from Tableau Desktop, data source: OpenDataPhilly, School Facilities [292]
Congrats, you've made it to the end of the course! In this lesson, we discussed how the recent re-visioning of the map reader as a map user has changed the cartographic design process, as well as how we evaluate maps. We discussed many different elements that may be integrated with maps, such as graphs, charts, and explanatory text, and explored the different mediums (e.g., interactive dashboards, data journalism) in which these elements are combined.
At the end of the lesson, we discussed when not to map, encouraging a practical approach to data visualization that views maps as a valuable tool but not a panacea. Relatedly, we note that much of cartographic design theory is widely applicable, and can be applied when designing other data visualizations or writing graphic-adjacent text—from microcontent to full articles.
In Lab 9, we designed an interactive map-based story using the visual analytics platform Tableau. Though this lab focused heavily on concepts from Lessons 8 and 9, we also drew from concepts throughout the course—refining layouts, symbolizing data, and thinking critically about map audience and purpose. This Story is now available on the web for you to share with others as demonstration of your skills in map design and data visualization. You're now ready and able to create, analyze, critique, and share high-quality maps!
You have reached the end of Lesson 9! Double-check the to-do list on the Lesson 9 Overview page [293] to make sure you have completed all of the activities listed there. After that, you'll have finished the course!
Links
[1] https://blogs.loc.gov/maps/2017/10/new-guide-to-russian-civil-war-pictorial-maps/
[2] https://news.psu.edu/story/489711/2017/10/23/research/sea-level-rise-not-stronger-storm-surge-will-cause-future-nyc
[3] https://esdac.jrc.ec.europa.eu/resource-type/national-soil-maps-eudasm?page=219
[4] http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2011:330:0039:0042:EN:PDF
[5] https://www.noaa.gov/
[6] https://daac.ornl.gov/ABOVE/guides/Northern_Alaska_Veg_Maps.html
[7] https://www.fairfaxva.gov/government/community-development-planning/zoning/zoning-subdivision-amendments/zoning-map
[8] https://creativecommons.org/licenses/by-sa/2.0/
[9] https://www.census.gov/library/visualizations/2016/comm/cb16-158_median_hh_income_map.html
[10] http://www.charts.noaa.gov/OnLineViewer/14774.shtml
[11] https://osfm.fire.ca.gov/what-we-do/community-wildfire-preparedness-and-mitigation/fire-hazard-severity-zones/fire-hazard-severity-zones-maps-2022
[12] http://JoshuaStevens.net
[13] http://www.marriott.com/hotels/event-planning/business-meeting/wasdt-washington-marriott-wardman-park/#
[14] https://blog.mass.gov/transportation/mbta/fare-hearings/
[15] https://www.beavertonoregon.gov/2311/Utility-Maps
[16] https://www.census.gov/geographies/mapping-files/time-series/geo/tiger-line-file.html
[17] https://www.atsb.gov.au/publications/investigation_reports/2013/rair/ro-2013-009
[18] https://sdotblog.seattle.gov/2017/12/06/transportation-is-the-key/
[19] https://www.e-education.psu.edu/geog486/sites/www.e-education.psu.edu.geog486/files/Lesson_01/Images/Fig1.5.5_0.jpg
[20] https://www.mapbox.com/gallery/
[21] https://www.mapbox.com/designer-maps/
[22] https://www.esri.com/arcgis-blog/products/arcgis-online/mapping/choosing-the-best-arcgis-online-basemap-for-your-maps-and-apps/
[23] https://www.usgs.gov/core-science-systems/national-geospatial-program/national-map
[24] https://catalog.data.gov/dataset/boundaries-zoning-districts-current
[25] https://www.mrlc.gov/
[26] https://www.axismaps.com/guide/general/visual-variables/
[27] https://creativecommons.org/licenses/by-nc-sa/4.0/
[28] http://downloads.esri.com/kfmaps/Sourcesandresources_poster.pdf
[29] https://archive.nytimes.com/www.nytimes.com/interactive/2011/05/01/weekinreview/01safe.html?_r=0
[30] https://twitter.com/NatGeoMaps/status/1082706163836375040
[31] https://www.nps.gov/grca/planyourvisit/upload/grca_pano.pdf
[32] https://pennstateoffice365.sharepoint.com/:u:/s/GEOG486FIles/EbCbBNLvOy5Lia_YOA55tcwBbMuwZX11Wj_htbh5W5xO-g
[33] https://www.usgs.gov/the-national-map-data-delivery
[34] https://nationalmap.gov/
[35] https://viewer.nationalmap.gov/basic/
[36] https://viewer.nationalmap.gov/services/
[37] https://www.mrlc.gov/index.php
[38] http://www.naturalearthdata.com/
[39] http://www.naturalearthdata.com/downloads/
[40] https://www.census.gov/geo/maps-data/
[41] https://www.census.gov/geo/maps-data/data/tiger-cart-boundary.html
[42] https://www.openstreetmap.org/#map=5/38.007/-95.844
[43] http://www.pasda.psu.edu/
[44] http://wvgis.wvu.edu/data/data.php
[45] https://www.kansasgis.org/index.cfm
[46] https://www.e-education.psu.edu/geog486/node/808
[47] http://www.igu-icatoponymy.org/
[48] https://commons.wikimedia.org/w/index.php?curid=2138205
[49] http://thinkingwithtype.com/
[50] https://designshack.net/articles/typography/serif-vs-sans-serif-fonts-is-one-really-better-than-the-other/
[51] http://mentalfloss.com/article/82867/whats-difference-between-font-and-typeface
[52] https://blog.prototypr.io/to-choose-the-right-typeface-look-at-its-x-height-instead-d5ef0967d09c
[53] https://www.charlottenc.gov/files/sharedassets/cats/cats-images/cats-uptown-map.png
[54] https://commons.wikimedia.org/wiki/Main_Page
[55] https://www.axismaps.com/guide/general/labeling/
[56] http://pro.arcgis.com/en/pro-app/help/mapping/text/label-with-the-maplex-label-engine.htm
[57] https://www.e-education.psu.edu/geog486/node/809
[58] https://www.census.gov/programs-surveys/geography/data/interactive-maps.html
[59] https://city.milwaukee.gov/
[60] https://www.nps.gov/blri/planyourvisit/rocky-knob-trails.htm
[61] https://www.health.act.gov.au/hospitals-and-health-centres/canberra-hospital/your-time-hospital/coming-hospital
[62] https://www.usgs.gov/media/images/image-2017-us-topo-map-jackson-wy-showing-standard-topo-layers
[63] https://www.pmel.noaa.gov/elnino/what-is-el-nino
[64] https://www.hayward-ca.gov/documents/downtown-specific-plan-existing-zoning-map
[65] http://ncei.noaa.gov/
[66] https://www.tandfonline.com/doi/full/10.1080/2373566X.2021.1883455
[67] https://www.nps.gov/subjects/cherryblossom/maps.htm
[68] https://www.nationalgeographic.org/media/atlas-happiness/
[69] https://www.census.gov/library/visualizations/2016/comm/cb16-158_poverty_map.html
[70] https://www.census.gov/dataviz/visualizations/001/
[71] https://www.census.gov/dataviz/visualizations/039/
[72] https://www.nrcs.usda.gov/
[73] https://www.census.gov/library/visualizations/2018/comm/population-change-2017-2018.html
[74] https://planthardiness.ars.usda.gov/
[75] https://www.census.gov/library/visualizations/interactive/big-game-census.html
[76] https://somethingaboutmaps.wordpress.com/2021/02/02/critique-with-empathy/
[77] https://proceedings.esri.com/library/userconf/proc13/papers/1015_20.pdf
[78] https://community.canvaslms.com/docs/DOC-10550-4212103951
[79] https://community.canvaslms.com/docs/DOC-10651-421254363
[80] https://pennstateoffice365.sharepoint.com/:u:/s/GEOG486FIles/EVUrdA6nUI5Jsy83nbrNfY4B-U-36tSrWlPjUw7wCcyjFA
[81] https://www.census.gov/cgi-bin/geo/shapefiles/index.php
[82] https://data.census.gov/cedsci/
[83] https://www.e-education.psu.edu/geog486/node/810
[84] http://colorbrewer2.org/
[85] https://earthobservatory.nasa.gov/blogs/elegantfigures/page/2/
[86] https://www.nist.gov/
[87] https://www1.nyc.gov/site/planning/data-maps/open-data/dwn-gis-zoning.page
[88] https://99percentinvisible.org/article/color-sphere-professors-pivotal-color-space-numbering-system/
[89] https://earthobservatory.nasa.gov/blogs/elegantfigures/2013/08/05/subtleties-of-color-part-1-of-6/
[90] https://www.nga.gov/conservation/science/chemical-characterization-of-pigments-in-art.html
[91] https://commons.wikimedia.org/wiki/User:Datumizer
[92] https://creativecommons.org/licenses/by-sa/3.0/deed.en
[93] https://www.census.gov/library/visualizations/2019/comm/irish-american.html
[94] https://earthobservatory.nasa.gov/blogs/elegantfigures/2013/08/12/subtleties-of-color-part-3-of-6/
[95] https://nca2014.globalchange.gov/sites/report/files/images/web-large/Figure-4.4-hi.jpg
[96] https://www.ncdc.noaa.gov/temp-and-precip/us-maps/
[97] https://www.census.gov/
[98] https://www.susielu.com/data-viz/viz-palette
[99] https://creativecommons.org/licenses/by-sa/4.0//
[100] https://www.illusionsindex.org/ir/20-checkershadow
[101] https://www.illusionsindex.org/illusions
[102] http://census.gov
[103] http://stats.org.uk/statistical-inference/TverskyKahneman1971.pdf
[104] https://www.phila.gov/departments/department-of-public-health/
[105] https://www.maine.gov/dacf/mfs/projects/fall_foliage/report/index.shtml
[106] https://hbr.org/2014/04/the-right-colors-make-data-easier-to-read
[107] https://pennstateoffice365.sharepoint.com/:u:/s/GEOG486FIles/EfDvs2ULYCJFg_5nfQmMDVoBx2JTk64AHGwnBYZap_V5yQ
[108] https://www.e-education.psu.edu/geog486/node/811
[109] https://www.utpjournals.press/doi/pdf/10.3138/9821-M648-7189-0088
[110] https://www.nasa.gov
[111] https://www.usgs.gov/
[112] https://www.jpl.nasa.gov/spaceimages/details.php?id=PIA12104
[113] https://www.icsm.gov.au/education/fundamentals-mapping/datums
[114] https://www.ngs.noaa.gov/datums/newdatums/index.shtml
[115] https://gistbok.ucgis.org/bok-topics/earths-shape-sea-level-and-geoid
[116] https://gistbok.ucgis.org/bok-topics/vertical-geopotential-datums
[117] http://gistbok.ucgis.org/bok-topics/map-projections
[118] https://oceanservice.noaa.gov/education/tutorial_geodesy/
[119] https://commons.wikimedia.org/wiki/File:Cilinderprojectie-constructie.jpg
[120] https://www.icsm.gov.au/education/fundamentals-mapping/projections
[121] https://www.vox.com/world/2016/12/2/13817712/map-projection-mercator-globe.
[122] https://thetruesize.com/
[123] https://pubs.er.usgs.gov/publication/70047422
[124] https://www.giss.nasa.gov/tools/gprojector/help/projections/GoodeHomolosine-Interrupted.png
[125] https://www.axismaps.com/guide/general/map-projections/.
[126] http://projectionwizard.org
[127] https://www.jasondavies.com/maps/transition/
[128] https://gist.github.com/d3indepth/f7ece0ab9a3df06a8cecd2c0e33e54ef
[129] https://bramus.github.io/mercator-puzzle-redux/
[130] http://www.radicalcartography.net/projections/state%20planes.gif
[131] https://pubs.usgs.gov/fs/2001/0077/report.pdf
[132] https://www.wired.com/2013/11/projection-lambert-conformal-conic/
[133] https://www.census.gov/dataviz/visualizations/051/
[134] https://www.e-education.psu.edu/geog486/sites/www.e-education.psu.edu.geog486/files/Lesson_05/Files/FreightMP080515%20%282%29.pdf
[135] https://creativecommons.org/licenses/by-sa/3.0
[136] https://upload.wikimedia.org/wikipedia/commons/5/5f/Minard%27s_Map_%28vectorized%29.svg
[137] https://creativecommons.org/licenses/by-sa/4.0
[138] https://upload.wikimedia.org/wikipedia/commons/e/e2/Minard_Update.png
[139] http://www.martingrandjean.ch/historical-data-visualization-mapping-migration-in-1862/
[140] https://www.nasa.gov/directorates/spacetech/spinoff/feature/space_radio_flight_tracking
[141] https://www.e-education.psu.edu/geog486/sites/www.e-education.psu.edu.geog486/files/Lesson_05/Files/Geog486_Lab5_New.zip
[142] https://www.flightradar24.com/
[143] http://pro.arcgis.com/en/pro-app/tool-reference/data-management/xy-to-line.htm
[144] https://www.e-education.psu.edu/geog486/node/812
[145] https://www.srs.fs.usda.gov/cradle/climate/index.html
[146] https://www.nps.gov/media/photo/view.htm?id=2C5A97C9-1DD8-B71C-0762857084D025D1
[147] https://www.loc.gov/maps/?q=Heinrich+Berann
[148] https://mars.nasa.gov/resources/3175/terrain-model-of-mars-mojave-crater/
[149] https://www.usgs.gov/news/mapping-national-parks-0
[150] https://www.loc.gov/item/75694832/
[151] https://landsat.gsfc.nasa.gov/wp-content/uploads/2018/05/20180508_Wrangell_mts_lg.png
[152] https://eros.usgs.gov/doi-remote-sensing-activities/2016/usgs
[153] http://www.shadedrelief.com/
[154] http://www.reliefshading.com/
[155] https://nasa3d.arc.nasa.gov/detail/moon-nearside-farside
[156] https://www.youtube.com/watch?v=CE1B7tdGCw0&feature=youtu.be
[157] https://www.youtube.com/channel/UCPpw6jRIL1B9nU0k7-GujOQ
[158] https://commons.wikimedia.org/wiki/Category:Contour_lines
[159] https://www.icsm.gov.au/education/fundamentals-mapping/types-maps/topographic-maps
[160] https://www.nps.gov/gaar/planyourvisit/maps.htm
[161] https://oceanexplorer.noaa.gov/explorations/02hudson/logs/sep01/media/contourmap.html
[162] https://gis.utah.gov/dumbing-down-terrain-with-contour-polygons/
[163] https://www.loc.gov/item/88693037/
[164] https://www.usgs.gov/media/images/shaded-relief-map-0
[165] https://unsplash.com/photos/Ak5c5VTch5E
[166] https://icaci.org/files/documents/ICC_proceedings/ICC2009/html/nonref/28_8.pdf
[167] https://www.esri.com/arcgis-blog/products/arcgis-pro/3d-gis/hacking-a-dem-sunrise/.
[168] https://ngmdb.usgs.gov/img4/ht_icons/Browse/NY/NY_Central%20Park_137165_1995_24000.jpg
[169] https://prd-tnm.s3.amazonaws.com/StagedProducts/Maps/HistoricalTopo/PDF/NY/24000/NY_Central%20Park_137165_1995_24000_geo.pdf
[170] https://pubs.usgs.gov/imap/i2781/
[171] https://pubs.usgs.gov/imap/i2781/i2781_c.pdf
[172] https://montrails.org/wp-content/uploads/2019/11/MRTC_Brochure2016_FINAL_Web-1.pdf
[173] https://pennstateoffice365.sharepoint.com/:u:/s/GEOG486FIles/ESyo3aQObdxOtYh52VYkxwsBNWNsmCkhZ5SQYc7pFVwe2w
[174] http://pro.arcgis.com/en/pro-app/tool-reference/3d-analyst/interpolate-shape.htm
[175] http://pro.arcgis.com/en/pro-app/help/editing/a-quick-tour-of-editing.htm
[176] https://www.e-education.psu.edu/geog486/node/813
[177] https://www.oregonmetro.gov/councilor/sam-chase/news/5587
[178] https://www.census.gov/dataviz/visualizations/062/
[179] https://www.axismaps.com/guide/multivariate/bivariate-choropleth/
[180] http://www.joshuastevens.net/cartography/make-a-bivariate-choropleth-map/
[181] https://www.nsf.gov/news/news_images.jsp?cntn_id=137907&org=NSF
[182] https://www.lee.senate.gov/public/index.cfm?p=socialcapitalproject&id=5DBFEE4E-D95A-4307-9909-4B687DBD3CC0
[183] https://www.axismaps.com/guide/multivariate/value-by-alpha/
[184] https://commons.wikimedia.org/w/index.php?curid=2454562
[185] https://www.census.gov/dataviz/visualizations/009/
[186] http://www.oecdbetterlifeindex.org
[187] https://mapdesign.icaci.org/2014/12/mapcarte-353365-life-in-los-angeles-by-eugene-turner-1977/
[188] https://www.youtube.com/watch?v=CGTfBll_sas
[189] https://geography.wisc.edu/hazardouswaste/
[190] https://flowingdata.com/2017/10/19/mapping-happiness/
[191] https://www.esri.com/arcgis-blog/products/arcgis-pro/mapping/chernoff-faces/
[192] https://www.census.gov/dataviz/visualizations/048/
[193] https://commons.wikimedia.org/wiki/File:KMeans-Gaussian-data.svg
[194] https://creativecommons.org/licenses/by-sa/3.0/
[195] http://pro.arcgis.com/en/pro-app/tool-reference/spatial-statistics/how-multivariate-clustering-works.htm
[196] https://www.sciencedirect.com/science/article/pii/S0167865509002323
[197] https://gistbok.ucgis.org/bok-topics/representing-uncertainty
[198] https://www.giss.nasa.gov/tools/panoply/overlays/
[199] https://www.climate.gov/news-features/featured-images/future-temperature-and-precipitation-change-colorado
[200] https://fortmillsc.gov/
[201] https://s3.amazonaws.com/happiness-report/2018/WHR_web.pdf
[202] https://www.un.org/sustainabledevelopment/sustainable-development-goals/
[203] https://pennstateoffice365.sharepoint.com/:u:/s/GEOG486FIles/ER6wQkPrAY5MosaUH_eX6wwBSGvr2rknl-72xfB8B2qZUw
[204] https://pennstateoffice365.sharepoint.com/:w:/s/GEOG486FIles/EVII8Kc9EeBMmxl5d9INvUgBS0ukJpHAV-Su0H6Zp0n_Tg
[205] http://worldhappiness.report/ed/2018/
[206] https://pennstateoffice365.sharepoint.com/:w:/s/GEOG486FIles/EVII8Kc9EeBMmxl5d9INvUgBS0ukJpHAV-Su0H6Zp0n_Tg?e=Shdz1u
[207] https://www.e-education.psu.edu/geog486/node/646
[208] https://www.naturalearthdata.com/
[209] http://naturalearthdata.com/
[210] http://MapShaper.org
[211] https://www.openstreetmap.org/#map=6/39.402/-73.916
[212] https://www.openstreetmap.org/copyright
[213] http://www.cartographicperspectives.org/index.php/journal/article/view/cp68-roth-et-al/15
[214] https://creativecommons.org/licenses/by-nc-nd/3.0/
[215] https://carto.com/
[216] https://leafletjs.com/
[217] https://www.mapbox.com/?utm_source=chko&utm_medium=search&utm_content=MapboxGeneric&utm_campaign=CHKO-PR01-BR-Mapbox-USA-Exact&gclid=Cj0KCQiA14TjBRD_ARIsAOCmO9Zt1BRjCOnXQ4SeUSCY0jXpKIO9-76h7WIULF66hTD2aJ1pRwKAETcaAtPSEALw_wcB
[218] https://www.socialexplorer.com/
[219] https://graphical.weather.gov/xml/rest.php
[220] https://fws.gov.training
[221] https://www.openstreetmap.org/#map=16/41.7662/-72.6814
[222] https://gis.ny.gov/
[223] https://radar.weather.gov/
[224] https://blogs.cdc.gov/yourhealthyourenvironment/2018/03/19/tracking-network-data-spotlight-poisonings/
[225] https://www.climate.gov/news-features/featured-images/what-will-average-us-temperatures-look-future-octobers
[226] https://blogs.loc.gov/maps/2017/04/places-in-american-civil-war-history-preparation-for-war/
[227] https://www.nytimes.com/2013/01/07/nyregion/after-pinpointing-gun-owners-journal-news-is-a-target.html?pagewanted=al
[228] https://studentaffairs.psu.edu/involvement-student-life/student-services/student-news-readership-program
[229] https://fivethirtyeight.com/wp-content/uploads/2016/10/silver-electionupdate-womenvoted.png?w=1150
[230] https://twitter.com/jerry_shannon/status/893117028671913984
[231] https://www.youtube.com/watch?v=zl26Gg_bzeA
[232] https://www.flickr.com/photos/walkingsf/5159866800
[233] https://creativecommons.org/licenses/by/2.0/
[234] https://static.makeuseof.com/wp-content/uploads/2018/07/apple_maps_route2.jpg
[235] https://www.plumplot.co.uk/goodies.html
[236] https://medium.com/multiple-views-visualization-research-explained/how-do-we-know-when-a-visualization-is-good-c894b5194b62
[237] https://www.youtube.com/watch?v=1Oqm6eO6deU
[238] https://www.flickr.com/photos/inl/32612549026
[239] http://sites.psu.edu/archaeology/
[240] http://sites.google.com/a/pdx.edu/visualizing-forest-futures/home
[241] https://sites.google.com/a/pdx.edu/visualizing-forest-futures/home
[242] https://news.psu.edu/story/501271/2018/02/23/campus-life/augmented-reality-app-reveals-campus-monuments-history-teaching
[243] https://www.flickr.com/photos/caseorganic/4336163331
[244] https://creativecommons.org/licenses/by-nc/2.0/
[245] https://vimeo.com/121523850
[246] https://vimeo.com/patriciadobrindt
[247] https://vimeo.com
[248] https://www.mapbox.com/mapbox-studio/
[249] https://www.mapbox.com/gallery/#map-14 via @mapbox
[250] https://www.mapbox.com/gallery/#map-19
[251] http://www.mapbox.com/mapbox-studio/
[252] https://docs.mapbox.com/vector-tiles/reference/mapbox-streets-v8/#road
[253] http://fontspace.com
[254] https://fonts.google.com/
[255] https://docs.mapbox.com/help/tutorials/create-a-custom-style/?utm_source=explorer-onboard1&utm_medium=email&utm_content=create-custom-style&utm_campaign=explorer-onboard1
[256] https://www.mapbox.com/videos/how-to/filter-what-data-appears-on-your-map-in-mapbox-studio/
[257] https://www.youtube.com/playlist?list=PLOlUoOtyTOXh8bwQ6MM7p7l6gfo4j8hch
[258] https://www.e-education.psu.edu/geog486/node/815
[259] http://www.nytimes.com/interactive/2018/us/2018-year-in-graphics.html
[260] http://www.washingtonpost.com/graphics/2018/ns/best-graphics/?noredirect=on&utm_term=.2311ffb20fa2
[261] http://www.nationalgeographic.com/2018/12/most-memorable-maps-and-graphics-of-2018/
[262] https://www.portlandoregon.gov/auditservices/article/692191
[263] http://www.nngroup.com/articles/ten-usability-heuristics/
[264] https://www.nngroup.com/articles/ten-usability-heuristics/
[265] http://blogs.cdc.gov/yourhealthyourenvironment/2018/03/19/tracking-network-data-spotlight-poisonings/
[266] https://www.usability.gov/how-to-and-tools/methods/prototyping.html
[267] https://snohomishcountywa.gov/236/Community-Crime-Map
[268] https://www.e-education.psu.edu/geog486/node/4
[269] https://www.researchgate.net/figure/The-Pennsylvania-Cancer-Atlas-viewable-on-the-web-at_fig1_5231133
[270] https://gis.cdc.gov/grasp/fluview/fluportaldashboard.html
[271] https://vimeo.com/30781990
[272] https://vimeo.com/jsndyks
[273] https://earthobservatory.nasa.gov/images/144351/the-seasonal-rhythms-of-ammonia
[274] https://www.washingtonpost.com/graphics/2018/ns/best-graphics/?noredirect=on&utm_term=.2311ffb20fa2
[275] https://www.nytimes.com/interactive/2018/us/2018-year-in-graphics.html
[276] https://www.nationalgeographic.com/2018/12/most-memorable-maps-and-graphics-of-2018/
[277] https://www.portlandoregon.gov/auditservices/article/692191#mapoflistings
[278] https://www.nngroup.com/articles/microcontent-how-to-write-headlines-page-titles-and-subject-lines/
[279] https://www.phila.gov/programs/combating-the-opioid-epidemic/reports-and-data/opioid-misuse-and-overdose-data/
[280] https://oklahomaworks.gov/workforce-economic-development-data-dashboards/
[281] https://public.tableau.com/profile/cary.anderson#!/vizhome/Lab9_Part1_Tutorial/Story1
[282] https://public.tableau.com/profile/cary.anderson#!/vizhome/Philly_Schools/Story1
[283] https://www.lynda.com/Tableau-tutorials/Tableau-10-Essential-Training/500540-2.html
[284] https://onlinehelp.tableau.com/current/pro/desktop/en-us/maps_howto_simple.htm
[285] https://onlinehelp.tableau.com/current/pro/desktop/en-us/maps_options.htm
[286] https://pennstateoffice365.sharepoint.com/:x:/s/GEOG486FIles/EaAD-Y1tRtBKnns9fJZfCOAB7CFXkenb2d_Ud9ob-NVCug
[287] https://help.tableau.com/v2020.3/pro/desktop/en-us/extracting_data.htm
[288] http://colorbrewer2.org/#type=sequential&scheme=BuGn&n=3
[289] https://carto.com/carto-colors/
[290] https://onlinehelp.tableau.com/current/pro/desktop/en-us/formatting_create_custom_colors.htm
[291] https://public.tableau.com/views/Philly_Schools/Story1?:embed=y&:display_count=yes
[292] https://opendataphilly.org/datasets/schools/
[293] https://www.e-education.psu.edu/geog486/node/816