Total: 1
We have developed and tested a spatial scan statistic for categorical, functional data (CFSS) - a data structure within which current approaches cannot identify spatial clusters. Our methodology combines an encoding scheme for categorical, functional observations with a nonparametric scan statistic. In a simulation study with three distinct scenarios, the CFSS accurately recovered the simulated spatial clusters and gave very low false positive rates, high true positive rates, and high positive predictive values. We have also used the CFSS to identify and characterize spatial clusters in French air pollution data from the winter of 2024.