The paper suggests and outlines a research program conceptual and empirical, the AI-Driven Housing Market Analytics: Predicting Affordability, Demand Shifts, and Urban Resilience, to U.S. cities. We contend that the current processes of housing production and consumption are products of interaction between economic, demographic, and environmental processes, the nonlinearities, spatial heterogeneity, and temporal volatility of which is a source of necessity but inadequacy of traditional approaches. Predictive analytics and machine learning (ML) can complement more conventional econometric methods by identifying more complicated trends in high-frequency, multi-source data (housing sales and rents, administrative records, mobility traces, climate hazards layers, and others). Simultaneously, AI techniques present data-bias, fairness, and governance issues, which need to be defined. This introduction specifies and clarifies research objectives, constructs specific research questions and testable hypotheses, and contextualizes the study in contemporary policy and scientific discussions- associating the recent evidence about the increase of housing costs burden, shortage of affordable housing, risks of sea-level and flood, and improvement of ML applications to estimate valuation and mobility. The important contributions will be as follows; (1) an integrated methodological protocol of interpretable ML, hybrid simulation (agent-based models coupled with ML), and scenario-based climate projections; (2) empirical tests at city level across typologies (high-growth, climate-vulnerable, post-industrial, and emerging mid-sized cities); and (3) a clear fairness and governance framework of housing AI. This study will aim at generating practical early warning signals of affordability stress, predicting probable demand-shift trajectories, and informing robust and equitable policy reactions.
Ebunoluwa Olamidotun Adejumo (Fri,) studied this question.