Where'd it go? How Geographic and Force-directed Layouts Affect Network Task Performance

Abstract

When visualizing geospatial network data, it is possible to position nodes according to their geographic locations or to position nodes using standard network layout algorithms that ignore geographic location. Such data is increasingly common in interactive displays of Internet-connected sensor data, but network layouts that ignore geographic location data are rarely employed. We conduct a user experiment to compare the effects of geographic and force-directed network layouts on three common network tasks: locating a node, determining the path length between two nodes, and comparing the degree of two nodes. We found a geographic layout was superior for locating a node but inferior for determining the path length between two nodes. The two layouts performed similarly when participants compared the degree of two nodes. We also tested a relaxed- or pseudogeographic layout created with multidimensional scaling and found it performed as well or better than the pure geographic layout on all tasks but remained inferior to the force-directed layout for the path-length task. We suggest interactive displays of geospatial network data allow viewers to switch between geographic and force-directed layouts, although further research is needed to understand the extent to which viewers are able to choose the most appropriate layout for a given task.