A national dataset of downtown business district boundaries for pre-automobile U.S. cities and towns

The downtown business district is a fundamental unit of analysis in urban revitalization research, yet no nationally consistent, publicly available dataset of downtown boundaries exists for small and mid-sized U.S. cities and towns. This paper introduces an open data product containing 6,551 downtown polygons for pre-automobile-era places spanning 47 contiguous states. The dataset is constructed from a fully reproducible, open-source R workflow: a national universe of historically significant places is assembled using historical census records, contemporary population data, and metropolitan classification criteria; points-of-interest data from the Overture Maps Foundation are collected within each place boundary; and a kernel density estimation algorithm identifies the highest-concentration commercial cluster as the downtown core. The resulting boundaries are validated against Main Street program districts, which the polygons cover at high rates in the two states examined. The polygons are deposited at Harvard Dataverse under a CC0 1.0 license, with all R code available on GitHub. The dataset supports research on downtown revitalization, Main Street commercial activity, and the economic geography of small and mid-sized American cities.