Abstract movers and their trajectories: movement characterization and uncertainty for aggregate spatial units.

Keywords: Spatiotemporal Clustering, Space-Time Prisms, Trajectories
Abstract Type: Paper Abstract

Abstract

Ubiquitous location aware technologies in science and society have produced an abundance of movement data describing the spatiotemporal arrangement of objects and events. In response, GIScientists and other researchers interested in pattern and process captured by movement data have developed an effective methodological literature. Proven techniques identify local spatiotemporal clusters of events, as well as bound and reconstruct the uncertain positions of objects in space at times between observation, respectively. The subject matter for these approaches center on the individual or atomic mover. The mover is often a person, animal, vehicle, etc. in space. Direct treatment of aggregate spatial processes or events as movers is understudied in the literature.
This paper borrows from local indicators of spatial autocorrelation and space-time prisms towards alternative strategies for the characterization of an “abstract mover” derived from spatiotemporal observations of areal or spatially aggregated phenomena. The goal is to provide measures valid for aggregate spatial units, having similar interpretation to those measures provided in GIScience for individual movers.
At regular temporal intervals, cluster membership of aggregate spatial units is determined and tracked. Cluster membership over future times is a function of persistence and adjacency among cluster members from prior times. Centrality within clusters is identified using a mean center or center of a standard deviational ellipse. Sequences of cluster centers comprise the trajectory for the resulting “abstract mover” object. Next, this research evaluates the application and interpretation of existing methods from movement science on the new trajectory, specifically the space-time prism and its derivatives.