Date of Award


Document Type

Open Access Thesis



First Advisor

Sarah E. Battersby


Animated choropleth maps allow for the compilation of potentially massive time-series datasets which can portray space-time change in a congruent manner. They are also becoming increasingly common for data visualization. When users view and interact with these maps, however, there is the likelihood that the human cognitive-perceptual system may be overwhelmed by a large number of simultaneous changes in each scene: this so-called `change blindness' is a common malady when viewing successive scenes, unless scene-to-scene graphical changes are salient enough to attract the fixation of the user. Even then, there may be a limit to the number of simultaneous changes that the user can perceive. This thesis examined the saliency of change occurring in map features by conducting a human-subjects study to explore the effect of intensity, number and pattern of change-clusters on a map user's ability to detect change. These characteristics can be quantified for a given animated choropleth map using a localized change metric, Magnitude of Change. This study found that, for generalized choropleth maps, clusters in which at least 80% of the polygons changed class were significantly more likely to be successfully detected than clusters with lower levels of class change; additionally, users performed more poorly with maps containing single clusters than for those with multiple clusters. There were no differences in accuracy for gender, or for whether or not the user played video games regularly, but domain expertise (i.e., having taken a prior geography class) had a positive effect on accuracy. It appears that, for maximum effectiveness, animated choropleth maps should consist of limited datasets, and be made simpler and more user-friendly.

Included in

Geography Commons