Historical Migration Regions and Present Regional Identities

Keywords: Migration region, Family tree, Community detection, Spatio-temporal network analysis
Abstract Type: Paper Abstract

Abstract

Regions are often defined as areas with common characteristics of physical, or human geography such as land cover, climate, language, and migration. Regions are often considered to be stable or unchanged in most geographical analyses, which is often not the case when the features that define regions are related with humans who constantly move and migrate. In this paper, we use a population-scale family tree dataset with its largest connected component containing 40 million individuals in one single tree to study migration over several centuries. We extract migration events by tracing children's birthplaces in each family and generate a state-to-state migration network in U.S. between 1789 and 1924. We partition the migration network into periods and apply community detection to identify migration regions, or in other words, network structures, and their evolution over a period of 135 years. Migration regions coincide with regional and national cultural forms such as foodways, dialects, and political divisions. Our results also reveal that migration regions evolved over time both changing their borders and their socio-demographic structure. We also found that some regions were formed by groups of areas that are geographically distant from each other, which we explain with Gold Rush and family ties, which attracted individuals to far away destinations. We think that identifying functional regions and their evolution using high resolution mobility data of today could help reveal processes such as neighborhood identities, and segregation at very different and small spatial and temporal scales than our study.