🤖 AI Summary
This study investigates spatial inequities in public transit accessibility to critical points of interest (POIs)—such as healthcare facilities and grocery stores—across major U.S. cities. We propose a multi-scale fairness analytical framework that integrates the two-step floating catchment area (2SFCA) method with interpretable machine learning, enabling cross-city comparison, stratified demographic assessment, and dynamic policy simulation of POI additions or removals. An interactive web platform supports accessibility visualization and scenario modeling. Empirical analysis reveals systematic spatial disparities in POI accessibility, disproportionately affecting low-income and racially minoritized communities. Notably, this work pioneers the integration of explainable AI (XAI) into the 2SFCA workflow, quantifying the marginal contributions of socioeconomic variables to accessibility gaps. The framework delivers a computationally tractable, simulatable, and intervention-aware decision-support paradigm for integrated transportation and public service planning.
📝 Abstract
The tool developed as a result of this paper analyzes the ease and equity of access to major POI categories (e.g. vaccination centers, grocery stores, hospitals) using public transit in major U.S. cities. We built an interactive website that enables easy exploration of current access equity and allows performing scenario analysis by introducing/removing POIs. Accessibility indices were calculated using a 2SFCA (2-step floating catchment area) approach, and ML methods were utilized for exploratory statistical analysis of the results.
