Visualizing large graphs by layering and bundling graph edges

Abstract

Edge bundling has been widely used to reduce visual clutter and reveal high-level edge patterns for large graphs. Due to strong edge attraction, bundled results often show unnecessary curvature and tangling at bundle intersections. Inappropriate bundling may fail to reveal true data patterns and even mislead users. This paper presents a parameterizable 6-step edge bundling approach called LEB that reveals the patterns of the input graph, with distinguishable and traceable bundles. The bundling results by LEB are also adjustable by tuning a small number of parameters. We have conducted a user experiment to test and compare LEB with previous approaches. The experiment on three datasets (including two common ones) demonstrates LEB’s superiority over previous approaches in visualizing data patterns. Our implementation with reusable computation also delivers an execution speed fast enough for real-time interaction and animation.