Hypenet: Visualizing Dynamic Hypergraphs

Abstract

We present Hypenet, a novel technique to visualize dynamic hypergraphs. Such structures can model multiple types of data, such as computer networks with multiple destination addresses (multicast) or co-authorship networks with multiple authors per article. Hypenet visualizes the evolving topology of the hypergraph in a compact way, allowing users to detect patterns and inconsistencies. We describe our technique and show how it applies to the case of the history of publications of the Eurovis conference, revealing interesting patterns that can contribute to tell a story about data and create hypotheses.