Content Importance Based Edge Bundling for Graph Visualization

Edge bundling is widely used to reduce visual clutter and enhance high-level structure of graph in visualization. However, in some cases edges that belonged to different structures may be bundled together, which resulted in decrease of readability. To solve this problem, this paper proposes an edge bundling algorithm based on content importance. Firstly, a relation measuring algorithm is used to obtain the information of each edge. Then, those edge clusters which can reflect the high-level structure of graph are explored from massive points and edges. Finally, an improved force-directed edge bundling algorithm is used to bundle those edge clusters. The edges with high importance are bundled into separate edge clusters on the premise of reducing visual clutter. Experiments show that, the algorithm can effectively reduce the influence between edge clusters that belonged to different structures in edge bundling, thus ensure connections in different structures to be properly bundled and high-level structure in graph displays more clearly.