GEOG 486
Cartography and Visualization

Integration of Maps and Information Graphics

PrintPrint

Although maps may be stand-alone products, they are often integrated with text, imagery (e.g., photographs) and information graphics within some larger product, such as a book, an atlas, a journal article, a periodical or a report. Although integration is easy enough to achieve in a technical sense by importing a JPEG into a word-processing document or even just copying and pasting several digital files together, thoughtful integration requires a bit more work. One of the main reasons for using different methods for representing information is because they can provide complementary views into a particular problem or phenomenon. Thoughtful integration requires consideration of what types of insights each representation may provide and what viewpoints may be useful for solving the problem at hand. In this concept gallery item, we'll concentrate on thinking about integrating maps and scatterplots, but the same principles of thinking about the value of each representation apply to any collection of visual representations.

Maps are primarily designed to record and communicate information about the spatial distribution of some phenomenon, and with some effort, we can use either pairs of single-variable maps or a single multivariable map to determine if there is some sort of geographical correlation between two phenomena (i.e., if there is any spatial similarity in their patterns). Scatterplots, on the other hand, can be used to easily and quickly investigate an attribute space or statistical correlations (i.e., the relationship of two or more attributes with each other), but tell us little about geographical correlations. We can improve our understanding of the phenomena of interest and their relationships with each other if we use both types of representations.

We can increase the level of integration between multiple types of representations by making symbolization choices that help map users more easily assimilate and relate the information they glean from each type of representation. For example, using the same color schemes and classification breaks in both representations and creating a legend that mimics the organization of the scatterplot (i.e., organizes the variables in the same order) may help the map user more easily compare the two representations (Monmonier, 1993) (See Figure 8.cg.10, below).

A bivariate map and a scatterplot work together to allow users to assimilate and relate the information offered by the two represtentations.
Figure 8.cg.10 In this example of a bivariate map and a scatterplot, we used the same class breaks and color scheme for each representation, allowing the map user to more easily make comparisons between the representations. For example, s/he might choose to look at a cluster in the scatterplot (e.g., all of the dark purple points - high cancer rates and high poverty rates) to see if these observations are geographically clustered or dispersed.

If you are able to use an interactive, digital format for the maps and graphics you are producing (e.g., as in maps that are constructed for an internet mapping service, a multimedia CD-ROM or an electronic atlas), you can also take advantage of more explicit integration techniques, such as linking the representations so that when observations are selected in one representation (e.g., the scatterplot), they are highlighted in all of the linked representations (see Figure 8.cg.11, below). This particular technique is called brushing, and is quite effective for determining whether clusters of points in the scatterplot cluster geographically or vice-versa (Monmonier 1989).

Example of a linked map and scatterplot.
Figure 8.cg.11 In this example, you can see a linked map and scatterplot. Linking is a particularly useful technique for investigating clumps of interesting observations to see if they cluster in both geographic space (i.e., in the map) and attribute space (i.e., in the scatterplot).

Recommended Readings

If you are interested in investigating this subject further, I recommend the following:

  • Monmonier, M. 1989. "Geographic brushing; enhancing exploratory analysis of the scatterplot matrix." Geographical Analysis. 21(4): 81-4.