Visualizing and Interacting with Time-Evolving Graphs

Large, time-evolving graphs can be directly and indirectly observed in a variety of phenomena across domains. Speci€cally, for following the evolution of communities over time, applications can range from tracking modules in protein-protein interaction networks [10] to groups in scienti€c co-authorship networks [3]. Other tasks may include anomaly detection and graph matching across graph snapshots in time. Whether these graphs are naturally present, as in friendships within a social network, or inferred, as in functional brain networks from fMRI data, pa‹ern discovery in data of this scale is non-trivial. Furthermore, as temporal graph analysis is of interest to a diverse set of disciplines, identifying an approach that is both relatively unbiased and interpretable is critical.