Urban Social Prediction

This project develops and validates machine-learning models for anticipating the social and economic consequences of urban renewal, and translates these models into interpretable planning tools.

About

Urban Social Prediction develops machine-learning models and decision-support tools for anticipating the social and economic consequences of urban renewal. The project addresses a practical planning problem: renewal decisions have long-lasting effects on housing, displacement, and community resilience, but these consequences are rarely visible to planners at the time decisions are made.

Urban renewal creates pressure on housing supply, infrastructure capacity, services, and social networks. It can bring investment and improved living conditions, but it can also produce displacement, unequal benefit distribution, and locally concentrated harm. This project uses data-driven methods to surface these dynamics before interventions are locked in.

Our technical work combines predictive modeling, spatial analysis, and interpretable interfaces designed for use by planning practitioners. We study which socio-spatial factors shape renewal outcomes, how models generalize across different urban contexts, and how predictions should be communicated to planning teams so they inform decisions without being mistaken for prescriptions.

The project also develops recommendation tools for strategic urban renewal planning. These tools help stakeholders compare parcel aggregation options, evaluate feasibility constraints, assess redevelopment potential, and reason about prioritization criteria. All outputs are designed to support deliberation rather than replace it — keeping social equity, local context, and planner judgment at the center of the process.