eCite Digital Repository

Evaluating esthetics for user-sketched layouts of clustered graphs with known clustering information


Lin, C-C and Huang, W and Liu, W-Y and Tanizar, S and Jhong, S-Y, Evaluating esthetics for user-sketched layouts of clustered graphs with known clustering information, Journal of Visual Languages and Computing, 37 pp. 1-11. ISSN 1045-926X (2016) [Refereed Article]

Copyright Statement

Copyright 2016 Elsevier Ltd.

DOI: doi:10.1016/j.jvlc.2016.09.001


This paper aims to empirically analyze the esthetics for user-sketched layouts of clustered graphs with known clustering information. In our experiments, given not only the adjacency list of a clustered graph but also its predefined clustering information, each participant was asked to manually sketch clustered graphs "nicely" from scratch on a tablet system using a stylus. Different from previous works, the main concern in this paper is on which graph drawing esthetics people favor when sketching their own drawings of clustered graphs with known clustering information. Another concern of this paper is on the esthetics of clustered graph layouts employed by participants which include not only characteristics and structures of the final graph layouts but also the behavior of user's sketching process (including layout creation and adjustment). By observing all layouts and drawing processes, the drawing strategies which participants applied and the drawing esthetics are analyzed. Results show that most participants were unsurprisingly able to draw graphs with clear presence of bridge edges and clustering cohesiveness; more importantly, to distinguish clusters within the restricted-size tablet screen during the drawing process, some of the participants were still able to make each cluster with fewer edge crossings, more symmetries, and more alignment of grid in a smaller drawing area where the cluster spreads. Our results support that to alleviate user's complex drawing tasks, aside from the grid-based editing function suggested by the previous work, graph drawing systems should also provide the clustering information if the structure of the graph to be drawn is known.

Item Details

Item Type:Refereed Article
Keywords:information visualization, graph drawing esthetics, clustered graph, visualization, user-sketched layout
Research Division:Information and Computing Sciences
Research Group:Library and information studies
Research Field:Human information interaction and retrieval
Objective Division:Expanding Knowledge
Objective Group:Expanding knowledge
Objective Field:Expanding knowledge in the information and computing sciences
UTAS Author:Huang, W (Dr Tony Huang)
ID Code:112919
Year Published:2016
Web of Science® Times Cited:9
Deposited By:Information and Communication Technology
Deposited On:2016-12-04
Last Modified:2017-12-18

Repository Staff Only: item control page