Visualising Moving Clusters using Cluster Flow Diagrams

Thompson, J. and Torabi, T.

Moving clusters represent groups of objects that move together, for instance, groups of people evacuating a building. However, because moving clusters are composed of lists of clusters, they are not directly interpretable by analysts in their raw form. Hence, this paper introduces `cluster flow diagrams', a clear, concise and aggregated visualisation of a collection of these clusters. In particular, cluster flow diagrams give a snapshot of all of the formations of moving clusters, the changes within them and their interrelationships with one another. Additionally, the clusters are characterised by their member's spatial movements through two composite visualisations. The diagrams are generated using a four stage method which includes provisions to minimise noise in the data and artifacts from the process. Lastly, cluster flow diagrams are generated and evaluated using a synthetic evacuation scenario, a gaze tracking experiment, and a collection of storm tracks, truck trips and naval vessel trajectories.

Cite as: Thompson, J. and Torabi, T. (2015). Visualising Moving Clusters using Cluster Flow Diagrams. In Proc. 38th Australasian Computer Science Conference (ACSC 2015) Sydney, Australia. CRPIT, 159. Parry, D. Eds., ACS. 37-45

(from crpit.com) (local if available)