Increasing the readability of graph drawings with centrality-based scaling

A common problem in visualising some networks is the presence of localised high density areas in an otherwise sparse graph. Applying common graph drawing algorithms on such networks can result in drawings that are not highly readable in the dense areas. Additionally, networks whose layouts are defined geographically often have dense areas that are located within small geographical regions relative to the size of the entire network. In cases where relationships within these dense areas are of interest, it is desirable to be able to distort the graph layout such that the denser areas are enlarged from their original sizes. In this paper, we propose a technique for enlarging dense areas of a given graph layout, and shrinking sparse areas. This technique is applied to geographical layouts of railway networks and force-directed layouts of non-geographical networks. The results show an increase in readability of dense parts of the networks. In addition, they provide improved starting layouts for schematisation methods which may be used to further increase readability.