NSF Award
2314726
This project investigates the interconnected role of amenities, residential preferences, mortgage lending practices, and real estate advertisements in neighborhood change processes. In this project, the researchers analyze how the language used in real estate advertisement text has evolved over time and varies by the race and income of anticipated neighborhood mortgage applicants. The use of real-time real estate listings at a point-level spatial resolution offers the potential to predict changes before they become too entrenched, enabling public policies to adapt timely. Finally, the project promotes public engagement and the use of science and technology in public policy by developing an online textbook for integrating natural language processing (NLP) in spatial analyses and by teaching K-12 girls enrolled in a STEM Camp about NLP methods and applications.<br/><br/>Housing market professionals, including realtors and mortgage lenders, have played a significant role in shaping neighborhoods by aiding in establishing and maintaining observed patterns of racial and income segregation across US cities. This project uses a combination of novel, theory-guided NLP, machine learning, and classic statistical methods to predict the racial and income composition of anticipated mortgage applicants in a neighborhood over time-based on the words used in property advertisements. It also investigates trends in mortgage denial rates as advertised housing and neighborhood amenities have shifted. Finally, the project develops new methodological approaches for examining housing dynamics at a fine spatial and temporal resolution.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
