| Abstrakt: |
Point pattern analysis is a fundamental analytical tool in various academic fields related to spatial information science. Basic but important patterns discussed in existing studies are spatially clustered and dispersed points. Analysis of these static patterns aims to reveal their underlying structure, i.e. why and how they are formed. However, to answer this research question, the direct tracking of the pattern formation process is more effective, though we have not yet fully conducted such analysis along the temporal axis. To fill the research gap, this paper develops a new method of spatiotemporal analysis. We focus on the point density around the appearance and disappearance of points. A research question is whether points appear/disappear in dense or sparse space. Extending Ripley's K-function, we develop four measures statistically evaluating the point density. We applied the proposed method to the analysis of the competition among convenience store chains in Shibuya-ku, Tokyo. A new method for event pattern analysis is proposed. The method evaluates the point density around the appearance and disappearance of points. The method considers event patterns within a single set of points and between two different sets of points. The method is applied to analyse the competition among convenience store chains in Shibuya-ku, Tokyo, which reveals the dynamic aspects of the competition among store chains. [ABSTRACT FROM AUTHOR] |
