EdgeAnalyzer: Exploratory Analysis through Advanced Edge Interaction

Abstract

Overdraw and edge crossings are common problems during the exploration of large graphs. Layout methods, such as edge bundling, have been applied to diminish this problem. They often come at the cost of introducing local ambiguities and losing connectivity detail, thus making edge interaction difficult. In this paper we propose EdgeAnalyzer, a focus+context technique for direct interaction with edges and edge bundles, keeping the original graph layout unaltered. Edges can be focused, grouped, filtered, and selected employing user-chosen criteria allowing for the analysis of the underlying data. EdgeAnalyzer can be applied to different kinds of node-link diagrams, but is particularly useful when similarities of edges can be exploited. We demonstrate EdgeAnalyzer with different example scenarios, and show the feasibility of the approach in more detail for co-classification analysis in the intellectual property domain.