Anthony Robinson and I designed and built this visualization for NACIS 2006. We had two goals that led to the designs techniques used in this visualization. The first was to find ways of building geovisualization tools that have the look of a printed atlas or clean PDF document but the feel of a fully interactive visualization user interface. The second was to make geovisual displays easier to understand and operate by eliminating unnecessary interface scaffolding and window "decoration", leaving behind the information-rich components of the display.

The visualization actually consists of two independent pages, each on its own tabbed-pane "page". The first page (screenshot above) shows a county-level map of the continental United States, with the the relationship between two selectable demographic variables as increasing green and purple coloring. In this case, the analyst is looking at the prevalence of health plans/insurance in different age groups, noting that retirees in the upper Midwest tend to be well covered.

The second page allows comparison of county-level results for the three main candidates running for President of the United States in 2004, using both mouseover highlighting and dynamic range filtering techniques to support exploration. (Download the live visualization and/or video tour to see both pages.)

In keeping with our design goals, interaction is simple. Health variables have been organized into a two-level hierarchy. A pair of lists on the left side allow selection of a category in the top level of the hierarchy, then one of its variables in the bottom level of the hierarchy. A univariate legend shows coloring for values of that variable, aggregated into five ranges using the natural breaks algorithm. An equivalent pair of lists and legend on the right side allow selection of a second health variable.

Mousing over counties in the map cause the corresponding cell in the legend to be highlighted. Conversely, mousing over bivariate breaks (color cells) in the legend cause corresponding counties to be highlighted in the map.

Visualization (2.3 MB)
Video Tour (14.4 MB)