On Clustered Graph Layouts Sketched by Users Based on Predefined Clusters

Abstract

The aesthetics for user-sketched layouts of clustered graphs with known clustering information is analyzed empirically. In our experiments, given the adjacency list of a clustered graph and its predefined clustering information, each participant was asked to manually sketch clustered graphs "nicely" from scratch on a tablet system using a stylus. The main feature of this work different from the previous work is to investigate if the user\'s complex task of sketching clustered graphs "nicely" can be alleviated by providing the cluster partition information. Results show that most participants can draw graphs with clear presence of bridge edges and clustering cohesiveness.