A Template for Interactive Conditioned Choropleth Maps

Keywords: Dynamic classes and conditioning, pan and zoom, interactions, cognostics

Abstract: The paper describes the template for conditioned choropleth(CC) maps. The paper extends the static coplots of Cleveland, Groose and Shyu (1992) to dynamic classed choropleth maps and emphasizes a 3 x 3 layout. The 2 x 2 layout with margins is an important special case. CC maps provide a simple way of thinking about three georeferenced variables, a dependent variable and two explanatory variables.

Interactivity facilitates exploration. Sliders control the selection of class boundaries and conditioning for each row and column. Regions satisfying constraints appear in foreground colors so only the color table needs to change. A small tab at the lower right of each panel provides a statistical summary, such as the combined mortality rate for highlighted regions. This provides a numeric check on main effects and interactions. Cognostics algorithms can help the analyst select interesting views. For closer views, a fixed aspect ratio (FAR) pan and zoom widget simultaneously controls all panels.

The examples show the 805 Health Service Areas as defined in the Atlas of United States Mortality. The mortality rates studied concern lung cancer and chronic obstructive pulmonary disease. CC maps are easy to use for beginning analysts and a powerful tool for sophisticated analysis seeking spatially linked explanation of regression residuals.