Abstract: Edge bundling has become a research hotspot in the field of information visualization. The
edge-bundling methods address the visual clutter problem caused by extensive edge crossings in graphs.
Among the recent edge-bundling methods, the algorithms which are based on the path construction are
generally efficient and have good bundling results, the algorithms which are based on edge clustering and
the skeleton construction can effectively reveal underlying patterns. Based on these works, a
distance-based edge-bundling method is presented, with the features of space distances and skeletons,
which can improve the edge-bundling results generated by the skeleton-based edge-bundling method.
The experiment results demonstrate that the distance-based method is efficient and effective in pattern
revealing, so that this method can avoid the over bundling problem of the previous one. In summary, this
method is a practical one that can simplify the computing process and avoid the singularity problem.

Key words: edge bundling, edge clustering, skeleton-based algorithm, graph visualization;
information visualization