Counterfactual mobility network embedding reveals prevalent accessibility gaps in U.S. cities

This paper investigates accessibility gaps in U.S. cities by analyzing large-scale mobility data. The authors develop a novel Counterfactual RANdom-walks-based Embedding (CRANE) method to learn continuous embedding vectors on urban mobility networks, disentangling confounding effects. Their analysis reveals significant income and racial gaps in mobility and visitation rates to certain venues, with bachelor’s degree holders experiencing greater mobility reduction during COVID-19. CRANE improves the explanatory capacity and robustness of revealed accessibility gaps, extending analysis to individual neighborhoods and enabling cross-city knowledge transfer. The findings highlight consistent accessibility gaps, calling for urban design policies to address these inequalities.

Yunke Zhang, Fengli Xu, Lin Chen, Yuan Yuan, James Evans, Luis Bettencourt, Yong Li