The Urban Institute created a multi-city panel dataset of land use reforms using machine learning algorithms to analyze newspaper articles across 8 metropolitan regions encompassing 1,136 cities from 2000-2019. The newspaper articles were accessed from Access World News, a comprehensive database of major national and international newspapers. We collected data on reforms related to 6 regulations: accessory dwelling units, density or floor-area-ratio requirements, minimum lot sizes, minimum setback requirements, and mixed residential and non-residential developments. For each reform, we classify whether it increased allowed residential density or decreased it and what time of regulation it fell under.

This dataset was produced with funding from the Smith Richardson Foundation.