Visual Analytics for Dynamically Conditioned Choropleth Maps: QQplots, Scatterplot Smoothes and Two-Way Tables
Chunling Zhang, (George Mason University), czhang1@gmu.edu,
Yaru Li, (George Mason University), yli6@gmu.edu, and
Daniel Carr, (George Mason University), dcarr@gmu.edu
Abstract
The visual analytics presented in this paper augment conditioned choropleth maps. In the conditioned map, dynamic sliders partition the map into a 3 x 3 grid of partial maps. Two different variables are attached to the two partitioning sliders. One slider controls row membership in the grid and their other controls column membership. The analyst's visual impression and comparison of the partial maps can be made more quantitative by showing other analytics. The analytics described in this talk are modifications of conventional QQplots, smoothed scatterplots, and two-way tables of means, effects, and model statistics. One modification involves the use of weights. Most modifications speed the response in order to keep up with the dynamic partitioning sliders. For example the smoothing widgets include the option to use an intermediate binning step when thousands of regions are involved. The talk provides live examples. The applications involve different kinds of region s! uch as county elementary school districts, hexagon grids for three states, and nations of the world.