Visually-supported graph traversals for exploratory analysis

Many real-world problems appearing in diverse application domains involve large multivariate interrelated data. For this reason, graphbased data models have gained popularity in recent years. Graph traversal is a powerful computational paradigm addressing the challenges of graph data management; yet, its complexity and specificity might hinder its use for interactive data exploration by non-expert users in absence of appropriate interfaces. We have designed and implemented a system for visually-supported graph traversal, featuring (1) a graphical block metaphor for traversal formulation and execution, and (2) data probes providing relevant visual feedback about the results. The proposed approach aims at enhancing the usability of graph querying and retrieval techniques, in order to assist users with gaining and interpreting insights during exploratory analysis.