Entry 1Bayesian Visual Analytics for Geo-Spatial Temporal Data
Feng Wang — Arizona State University

We explored the impact of test center utilization due to the addition or removal of testing sites. A spatial Voronoi partition of the data using all test center locations as seeds was created. The resultant partition has one test center per polygon and serves as a means of estimating center coverage. We decided to compare the effect that the removal or addition of a center would have from year to year. The change in utilization rate was calculated as:

where i is the month under analysis and j is the year under analysis. In this way, a 6 x 2 small multiple matrix was created where each cell represents the seasonal difference in center utilization (for example the difference between June 2014 and June 2013). Next, we filtered the data such that only Voronoi polygons that were within 25 miles of a center opening/closing compared to the previous season were rendered. Finally, connected Voronoi polygons were aggregated and our visualization shows the local average change in utilization rates due to the addition/removal of nearby centers. Regions in red represent a reduction in utilization and regions in blue represent an increase in utilization.

Feng Wang – fwang49@asu.edu, Brett Hansen-bjhanse1@asu.edu, Ross Maciejewski-rmacieje@asu.edu 
School of Computing, Informatics and Decision Systems Engineering
Arizona State University
699 S. Mill Ave
Tempe, AZ 85281