Interactive visualizations for decision support: Application of Rasmussen's abstraction-aggregation hierarchy.

ABSTRACT

Data visualization has of late received an enormous amount of attention from both researchers and practitioners. Even the popular press often includes impressive visualizations of various data sets. Interactive visualizations frequently include data visualizations, but they differ in that users employ the visualizations to make inferences, reach conclusions, and make decisions that result in changed and/or new visualizations. Data visualizations emphasize "what is," but interactive visualizations address "what if." In this way, interactive visualizations are often intended for decision support. This article addresses the design of interactive visualizations for decision support. An overall methodology is presented; central to this methodology is Jens Rasmussen's abstraction-aggregation hierarchy. The results of two applications and evaluations of the outcomes of using this methodology are discussed. The first application focused on interactive visualizations for helicopter maintenance. The second application addressed "enterprise diagnostics" in the automobile industry where subjects were asked to diagnose the cause of failed automobile brands. The results of these two applications are used to assess the efficacy of the proposed methodology.