Mapping migration regions and their evolution from population-scale family trees: what can they tell us about cultural identities and regions today?

Using a population-scale family tree dataset, this paper proposes a study of migration regions and their evolution in the U.S. between 1789 and 1924. To extract migration events, we use the child ladder approach, which traces family moves based on changes in birthplaces of consecutive children in each individual family. We calculate a time series measure of migration rate and partition the time into optimal periods so that each period has a distinct migration network. We apply community detection to derive migration regions from each network of different periods. We map these regions and use a pair-counting measure to statistically compare the similarity of regions in consecutive time periods. Migration regions reveal the extent to which the strong regional identities we see today, and, in the past, which were rooted in migration. The North/South divide was pervasive not only in the early periods but throughout U.S. history. Migration regions are important for understanding the development of regional and national cultural forms such as music, literature, foodways, and dialects, as well as political divisions and events.