Bivariate Spatial Clustering Analysis of Point Patterns: A Graph-Based Approach

Point pattern analysis is concerned with characterizing a spatial point process. A bivariate point process is one that generates points that are marked with binary values. There exists of dearth of methods for the spatial-analysis of non-numerical marked point pattern data, while these forms of data are increasingly common as a result of volunteered geographic information and geographically-indexed social media data. This paper highlights the problem of bivariate point clustering. A new method based on Delaunay triangulation is presented. Simulation studies are carried out to compare the new approach to existing methods. A case study examines clustering of antimicrobial resistance in Sri Lankan shrimp farms to illustrate the strengths and weaknesses of the method.