Interfaces and Visual Analytics for Visualizing Spatio-Temporal Data with Micromaps

INTERFACES AND VISUAL ANALYTICS FOR VISUALIZING SPATIOTEMPORAL DATA WITH MICROMAPS Chunling Zhang, Ph.D. George Mason University, 2012 Dissertation Director: Dr. Daniel B. Carr This dissertation addresses the visualization of spatio-temporal data for the purposes of communication, understanding, discovery, and analysis. A common approach to spatiotemporal visualization seeks to convey changes by showing a sequence of maps over time that does not address the problem of change blindness. The dissertation builds upon and extends the work of Carr and Pickle [2010] that calls attention to this problem and addresses visualization of spatial data patterns with conditioned and comparative micromaps. The dissertation adds dynamic interactivity to the comparative maps that address change blindness, develops new designs and demonstrates their use. Examples use the resulting interactive visualization tool named TCmaps (Temporal Change maps) to illustrate the designs and general utility by showing data from a variety of domains including health, education, environment, demography, and ecology. The examples show that one can see, point at and talk about all the changes designated by interactive thresholds since the changes are shown explicitly in separate maps sequences. The interactive designs address additional issues such as setting thresholds for categories whose changes are to be observed and methods for viewing change map sequences that are too long to see in one view. TCmaps is a general visual analytical tool. The dissertation emphasizes displaying results for two kinds of simulation models. One is a computational fluid dynamics model that simulates the transport and dispersion of toxic releases. The second shows the application of TCmaps to understanding the spread of an invasive species, the Eurasian CollaredDove (ECD) in the United States. In the last two decades researchers have developed many spatio-temporal models to characterize the invasion of ECD. This research modifies the hierarchical Bayesian matrix model developed by Hooten et al. [2006a] to provide new simulation results that better account for the species’ mysterious northwestern expansion. In this case study, TCmaps provides a powerful platform to display the spatial context of bird survey routes, observations, estimates, forecasts and variances. Since the research and methodology development in this dissertation builds on advances in the cognitive, data, statistics, and computing sciences, the results are of potential interest for a large domain of application.