Graph Layout with Versatile Boundary Constraints

Abstract

Graph layouts are in general data dependent and help to reveal structural and attribute relationships in the data set. However, there are situations when one may wish to alter the layout e.g. to emphasize parts of the data set or for aesthetic reasons. This paper strives to meet that need for the case of force-directed graph layout algorithms. Our approach is to add boundary constraints to specify where graph nodes may or may not be positioned. Users can interactively draw one or more boundaries. Boundaries may self-intersect and define different topology e.g. donut or figure eight shapes. Additional control, subject to the density of nodes, can impart different density distributions within defined boundaries. We tested the feasibility of this concept on several data sets and different boundary constraints.