Data visualizations are versatile tools for gaining cognitive access to large amounts of data and for making complex relationships in data understandable. This paper proposes a method for assessing data visualizations according to the purposes they fulfill in domain-specific data analysis settings. We introduce a framework that gets configured for a given analysis domain and allows to choose data visualizations in a methodically justified way, based on analysis questions that address different aspects of data to be analyzed. Based on the concepts addressed by the analysis questions, the framework provides systematic guidance for determining which data visualizations are able to serve which conceptual analysis interests. In a second step of the method, we propose to follow a data-driven approach and to experimentally compare alternative data visualizations for a particular analytical purpose. More specifically, we propose to use eye tracking to support justified decisions about which of the data visualizations selected with the help of the framework are most suitable for assessing the analysis domain in a cognitively efficient way. We demonstrate our approach of how to come from analytical purposes to data visualizations using the example domain of Process Modeling Behavior Analysis. The analyses are performed on the background of representative analysis questions from this domain.
In Software & Systems Modeling, vol. 19 (2020), pp. 531–554.