Recommendations for Data Visualizations Based on Gestalt Patterns

Abstract

An increasing amount of automated business processes and more intensive network communication among enterprise information systems leads to continuously growing amounts of data, which to understand requires to find cognitively adequate modes of representation. One is to use data visualizations. Support for efficiently selecting appropriate data visualizations based on specific information demands, however, is yet very limited. This article suggests a model infrastructure which allows to enrich syntactical matching patterns between data and visualization elements by associating Gestalt Patterns to both the characteristics of available data, and to visualization types. Based on these uniformly associated Gestalt Pattern characteristics, a distance measure can be computed between available data and available visualization types, which forms the basis for performing an automatic ranking of visualization types to support users in selecting visualizations appropriate to their information demands.