An Interactive Visual Language for Spatiotemporal Patterns

Abstract

Patterns exist in many contexts and they can be considered as knowledge sources providing useful information for our decision making. However, sometimes they are not easily identifiable and may not be visible. A visual interface for helping users to recognize patterns is needed. In this paper, we present a computer-based visual language called STVL to visualize spatiotemporal patterns. An event-anchored data set is processed to identify symbols of the language and those symbols are represented in a three-dimensional space consisting of latitude, longitude and time. The user then creates and modifies visual statements using interactive functions, which gives the dynamic syntax of the language. Visual reasoning is carried out in a cycle of interaction and meaningful visual statements are selected as spatiotemporal patterns. In this paper, we also present an application based on the STVL visual language, Pattern Browser, to demonstrate its applicability in information visualization.