A GeoAI-informed floating catchment area framework for assessing spatiotemporal equity of park accessibility for elderly population using dynamic popularity weighting and 3D pedestrian network

Source abstract

Equitable access to urban parks is an important determinant of healthy aging, yet conventional accessibility assessments frequently overlook temporal variability in park usage, terrain-induced walking impedance, and fine-grained demand heterogeneity among elderly populations. This study addresses several of these limitations by developing an integrated spatiotemporal analytical framework informed by Geospatial Artificial Intelligence (GeoAI) principles. The framework combines an enhanced two-step floating catchment area (E2SFCA) model with dynamic popularity weighting derived from Google Maps visitation data and a three-dimensional (3D) pedestrian network constructed from OpenStreetMap road geometries and ALOS PALSAR Radiometrically Terrain Corrected (RTC) data. The framework is applied to the high-density urban context of Macau characterized by rapid population aging, using building-level elderly population data from the 2021 census. A Spatially Tuned Poisson Regression (STPR) model is employed to empirically calibrate the distance-decay parameter and maximum service distance using revealed-preference review counts. Elderly population-weighted park accessibility is computed across six temporal scenarios defined by day type (weekday vs. weekend) and time period (morning, afternoon, evening). Equity is evaluated through Gini coefficients and the proportion of buildings falling below the city-wide median accessibility score. Results reveal pronounced spatial inequality in districts such as Barra / Manduco and Universidade e Baía de Pac On, which consistently exhibit accessibility deficits exceeding 95% of buildings below the median; while districts like Coloane, NAPE, and Cidade e Hipódromo da Taipa demonstrate comparatively favorable conditions. Temporal variation is modest yet notable, with morning periods generally associated with lower accessibility scores than afternoon and evening periods. The Gini coefficient analysis identifies Jardins do Oceano e Taipa Pequena and Coloane as the most internally unequal districts. Interpreted through the lenses of age-friendly city planning and environmental vulnerability, these findings offer an empirically grounded and temporally nuanced evidence base for prioritizing park infrastructure investments and targeted interventions to enhance equitable green space access for aging populations in compact urban environments.