Topological analysis of U.S. city demographics

Abstract

We apply persistent homology, the main method in topological data analysis, to the study of demographic data. Persistence diagrams efficiently summarize information about clusters or peaks in a region's demographic data. To illustrate how persistence diagrams can be used for exploratory analysis, we undertake a study of the 100 largest U.S.~cities and their Black and Hispanic populations. We use our method to find clusters in individual cities, determine which cities are outliers and why, measure and describe change in demographic patterns over time, and roughly categorize cities into distinct groups based on the topology of their demographics. Along the way, we highlight the advantages and disadvantages of persistence diagrams as a tool for analyzing geospatial data.