A Comprehensive Process for Spatiotemporal Analysis of Network-Based Phenomena




Eckley, David C

Journal Title

Journal ISSN

Volume Title



This thesis describes an efficient and effective process for conducting spatiotemporal cluster analysis of network-based phenomena. While various methods are published which describe spatiotemporal analysis of phenomena in continuous space, the literature is lacking for the application of these methods in network space. Through a step by step process, Part 1 of this thesis establishes the validity of a network application for the spatiotemporal clustering method proposed by Knox (1964). Further, it presents an intuitive technique for determining the critical parameters in space and time for the spatiotemporal test, by examination of minimum and average nearest neighbor distances in both dimensions independently. Through examples, Part 2 explores the significance tests used by the described spatiotemporal clustering methodology and expounds upon the critical parameter determination mentioned in Part 1. Part 3 presents a GIS-based toolbox, SCAn (Spatiotemporal Cluster Analysis on a network) designed to perform spatiotemporal cluster analysis of network-based phenomena using the methods presented in Part 1 and Part 2.



Spatiotemporal, Spatial Statistics, Cluster, Space-Time, Knox Test, GIS Software