Interactive analysis of big geospatial data with high‐performance computing: A case study of partisan segregation in the United States

Abstract Researchers are increasingly working with large geospatial datasets that contain hundreds of millions of records. At this scale, desktop GIS systems typically fall short and so new approaches and methods are needed. The objective of this work is to develop new approaches to interactively analyze large datasets and then to demonstrate the usefulness of those approaches using a case study looking at voter, or partisan segregation. Historically, the measurement of partisan segregation has been limited to comparing large geographic areas such as counties or states because researchers only had access to aggregated data. In this case study, however, we measure partisan segregation down to the individual for 180 million U.S. voters using advanced geospatial data science and high‐performance computing. This article discusses interactive method development for big geospatial data analysis including techniques used, solutions developed, and processing time statistics.