Patterns in Spatio-Temporal Data

Abstract

Spatio-temporal data is any information relating space and time. This entry specifically considers data involving point objects moving over time. The terms entity and trajectory will refer to such a point object and the representation of its movement, respectively. Movement patterns in such data refer to (salient) events and episodes expressed by a set of entities. In the case of moving animals, movement patterns can be viewed as the spatio-temporal expression of behaviours, as for example in flocking sheep or birds assembling for the seasonal migration. In a transportation context, a movement pattern could be a traffic jam. Only formalised patterns are detectable by algorithms. Hence, movement patterns are modelled as any arrangement of subtrajectories that can be sufficiently defined and formalised, see for example the patterns illustrated in Figure 1. A pattern usually involves a certain number of entities. It furthermore starts and ends at certain times (temporal footprint), and it might be restricted to a subset of space (spatial footprint).