GazeGraphVis : visual analytics of gaze behaviors at multiple graph levels for path tracing tasks

Abstract

Graph visualization contributes to an efficient understanding of interconnected properties in graph data. However, the exponential growth of interconnections poses great challenges to the efficient visual cognition of graph data. Generation of expressive graph visualization requires investigations of the cognitive process of exploring graph visualization, which can be revealed through the analysis of gaze behaviors. In this paper, we propose GazeGraphVis, a visual analytics system, to analyze gaze behaviors for path tracing tasks. Specifically, GazeGraphVis visualizes gaze behaviors at multiple levels of graph, node and edge with overview + detail techniques to provide a comprehensive analysis of human cognitive processes when finishing path tracing tasks. The insights of gaze behaviors for path tracing tasks are revealed using an integrated multiple-view interface. Domain experts in visualization and eye tracking analysis gave high praise to GazeGraphVis for its capability of obtaining the overall search tendencies and deeply analyzing the factors that affect gaze behaviors.