DGSVis: Visual Analysis of Hierarchical Snapshots in Dynamic Graph
Dynamic graph visualization attracts researchers' concentration as it represents time-varying relationships between entities in multiple domains (e.g., social media analysis, academic cooperation analysis, team sports analysis). Integrating visual analytic methods is consequential in presenting, comparing, and reviewing dynamic graphs. Even though dynamic graph visualization is developed for many years, how to effectively visualize large-scale and time-intensive dynamic graph data with subtle changes is still challenging for researchers. To provide an effective analysis method for this type of dynamic graph data, we propose a snapshot generation algorithm involving Human-In-Loop to help users divide the dynamic graphs into multi-granularity and hierarchical snapshots for further analysis. In addition, we design a visual analysis prototype system (DGSVis) to assist users in accessing the dynamic graph insights effectively. DGSVis integrates a graphical operation interface to help users generate snapshots visually and interactively. It is equipped with the overview and details for visualizing hierarchical snapshots of the dynamic graph data. To illustrate the usability and efficiency of our proposed methods for this type of dynamic graph data, we introduce two case studies based on basketball player networks in a competition. In addition, we conduct an evaluation and receive exciting feedback from experienced visualization experts.