Perception in Network Visualization

Abstract

Networks are used to model and represent data in many application areas from life sciences to social sciences. Visual network analysis is a crucial tool to improve the understanding of data sets and processes over many levels of complexity, such as different semantic, spatial and temporal granularities. While there is a great deal of work on the algorithmic aspects of network visualization and the computational complexity of the underlying problems, the role and limits of human perception are rarely explicitly investigated and taken into account when designing network visualizations. To address this issue, this Dagstuhl Seminar raised awareness in the network visualization community of the need for more extensive theoretical and empirical understanding of how people perceive and make sense of network visualizations and the significant potential for improving current solutions when perception-based strategies are employed. Likewise, the seminar increased awareness in the perception community that challenges in network research can drive new questions for perception research, for example, in identifying features and patterns in large, often time-varying networks. We brought together researchers from several different communities to initiate a dialogue, foster exchange, discuss the state of the art at this intersection and within the respective fields, identify promising research questions and directions, and start working on selected problems.