GEOG 588
Planning GIS for Emergency Management

Emerging Theme: Non-Expert Systems


Non-Expert Systems

A persistent issue with the application of GIS to emergency management is that there is often a significant knowledge gap between the GIS experts who know and understand the tools and intricacies of geographic data and the decision makers who have to act on information derived from GIS systems. Training decision makers on the use of GIS is not practical in most circumstances (and during a crisis situation, it's already too late).

This problem is not unique to GIS systems for emergency management. Businesses and organizations in many other domains experience the same situation. It is common enough now that an entirely new area of software development has emerged to meet this challenge, focused on the design and creation of so-called executive information systems that take the output of high-level analysis tools and translate that information into consumable and actionable intelligence geared toward the problems and level of technical understanding that decision makers have.


Recent years have seen a huge rise in executive information systems known as Dashboards. The idea behind dashboards is quite simple - what decision makers need most of the time is a strategically-oriented overview of their data. When we drive in a car, we have several basic information displays available that provide real-time feedback to help us control and guide the vehicle. As a result, many software dashboards resemble real world dashboards and industrial gauge panels to play off this common design metaphor (even though this is often a terrible idea).

Tableau Software offers tools to help analysts quickly assemble dashboards that are viewable and interactive through standard web browsers. Tableau focus on standard tabular datasets, but also include some basic mapping capabilities, and having used previous versions of their software myself, I recommend you check out what they offer if you think dashboards are in your organization's future (note - I have no stake in Tableau - they paid me zero dollars to say that). Here is a nice video overview of Tableau's tools that shows how data analysis tools connect to web-based dashboards.

Screenshot from a Tableau Dashboard showing a mix of business indicators in an easy to understand display.

Unfortunately, many dashboards (those from Tableau notwithstanding) are designed with flashy graphics that grossly violate best practices for the graphic display of data. It is enough of a problem now, that folks like Stephen Few are publishing books and working very nicely-paid consulting gigs to revamp bad dashboards to make sure their visual elements effectively communicate underlying data (take his Graph IQ test - it's good for anyone who works with maps and graphics to know this stuff).

GIS & Dashboards

Not surprisingly, GIS-driven dashboards are becoming more and more common. In many respects, recent work by Esri to promote and enhance ArcGIS Online and their StoryMaps templates could be considered examples of geospatial dashboard. I consider this a positive sign for the future of GIS in all types of organizations - for too long, what we do in GIS has been difficult to communicate to the outside world, and that has begun to change with things like widely-accessible web mapping tools and decision support dashboards. If we are able to design systems that reduce the time and steps required to communicate what is known from GIS analysis to decision makers then it becomes possible to more efficiently address the emergency situation. The fact that you can tweet out a live link to your semi-interactive map from ArcGIS Online is a really positive step.

Some design criteria for GIS-enabled dashboards could include:

  • The system must allow non-programmer analysts to quickly customize the dashboard (for example, when conditions change, data sources are modified, etc...).
  • Dashboard tools and graphics must be readily understandable by decision makers who have no GIS training.
  • Dashboards tools and graphics should focus on strategic objectives and the "big picture."
  • Time is an essential element - it has to be possible to understand changes over time in a disaster scenario.
Screenshot of SpatialKey
SpatialKey offers very clever map-enabled dashboards designed for non-experts to explore and analyze spatial data. This example shows forecasted impacts from Hurricane Sandy for use by an insurance compan.

Finally, I'd like you to take a look at a fascinating (crazy?) video that was developed by some folks in the visual analytics community (Information Visualization and Visual Analytics group at PNNL) who are charged with the task of envisioning new non-expert systems that help people collaborate in complex decision making scenarios. The concept is called Precision Information Environments:

For even more information on whether or not this stuff is really even possible, check out Precision Information's annotated browser version, which highlights how current emerging technologies directly inspired the things that are featured in this video.

Deliverables for this week's emerging theme

  • Post a comment on the Lesson 6 Emerging Theme Discussion in Canvas on how you could envision non-expert systems featured as part of a GIS system for emergency management. What are the challenges and opportunities here? Are there some aspects of emergency management that could never be made simple enough to appear in a dashboard? Are there other aspects that obviously "should" be visible in simple dashboards?
  • Then, I'd like you to offer additional insight, critique, a counter-example, or something else constructive in response to one of your colleagues' posts.
  • Brownie points for linking to other technology demos, pictures, blog posts, etc... that you've found to enrich your posts so that we may all benefit.