In many disciplines, data sets can be understood best by visualizing them as graphs, where the nodes describe the entities of interest, and the edges describe their relationships; in this way, for example, biologists can study the feeding relationships across organisms in an ecosystem. The problem is how to lay out large graphs in the plane (on a computer screen) in such a way that they can be conveniently investigated. This paper describes the interactive graph visualization software TreePlus, designed for this purpose.
The core idea of TreePlus is to use a tree layout that allows the user to interactively explore a graph by the metaphor of “plant a seed and watch it grow.” Starting with a specific node, the graph is expanded in layers: by first focusing on a node, all related nodes are shown in a preview (exhibiting those nodes that are new to the current view, as well as those that are already present). If this node is finally selected, the related nodes are added in a new layer (moving already present nodes to their new positions).
After discussing the TreePlus interface, the major part of the paper presents a controlled experiment that let participants perform specific tasks with TreePlus, and with another piece of software that uses a conventional graph layout. With TreePlus, the participants could solve many tasks faster; they also indicated that they felt more comfortable with the tree layout. Thus, the paper demonstrates how hierarchical presentations can aid humans in the understanding of data relationships.