From Hubs to Deserts: Urban Cultural Accessibility Patterns with Explainable AI

Cultural access is unequally distributed within cities: The research found a "pronounced core-periphery gradient" where museums, theaters, and galleries concentrate in urban centers, leaving peripheral neighborhoods—particularly lower-income areas—with significantly fewer cultural resources, though libraries provide more equitable coverage across density levels.

Explainable AI can reveal hidden patterns of inequality: By using interpretable machine learning models (SHAP), the researchers identified specific factors driving unequal access—income, population density, and racial/ethnic composition—making it possible to understand why disparities exist rather than just documenting that they do.

Open data and scalable methods enable cities to measure their own equity: The framework uses publicly available data and can be applied across different cities, giving urban planners and policymakers a practical tool to identify "cultural deserts" and track whether interventions actually improve access over time.

Income and race remain significant predictors of cultural access: The study found that non-library cultural infrastructure accessibility is "modestly higher in higher-income tracts," suggesting that systemic inequities in cultural opportunity persist and correlate with both economic and racial/ethnic patterns.

Libraries serve a different equity function than other cultural institutions: Unlike museums and theaters, library accessibility is "slightly higher in denser, lower-income areas," indicating that public libraries may be playing a more equitable role in cultural access—a finding that could inform funding and policy decisions.