Attribute-based edge bundling for visualizing social networks

• Similarity analysis of edges.
• Aggregation of similar nodes.
• Link-based edge bundling.
• Conduct satisfactory experiments.

Most nodes in complex networks have multiple attributes, which make them hard to analyze. Because general edge bundling algorithms fail to handle complex networks as a result of their intricate features, network simplification is extremely important. This paper proposes an attribute-based edge bundling algorithm that displays similar edges in nearby locations. Meanwhile, by analyzing complex networks at a community level, the overlapping clustering of nodes is well implemented, and better clustering effects can be achieved by grouping similar edges together. On the basis of datasets with different types and sizes, the experiments illustrate the simplification degree of the intricate graphs created by the algorithm proposed, which outperforms established competitors in correctness and effectiveness.