A Statistical Machine Learning Approach to Identify Rental Properties From Public Data Sources

Daniel Kuhlmann
University of Arizona

Jane Rongerude
Iowa State University

Lily Wang
George Mason University

GuanNan Wang
College of William and Mary

For academic researchers and practitioners alike, identifying individual rental properties can be incredibly useful but is often difficult due to insufficient and incomplete data. Although some cities have ordinances that require residential rental property owners (RRPOs) to register their properties, the availability and completeness of these registries vary dramatically from place to place. In places without rental registries, tax assessor data can provide some information but often not enough to clearly distinguish residential rentals from owner-occupied units and other commercial properties.