Growth of mapped park and fitness infrastructure in Latin American capital cities, with a weak income gradient

Introduction: Location-based augmented reality games send players to parks and public spaces, but open and comparable data on park infrastructure in Latin America are scarce. This study described the change in park and fitness infrastructure mapped in OpenStreetMap in Latin American capital cities and its distribution by income group. Methods: Ecological, repeated cross-sectional study of 20 capital cities, each defined as a 10-km-radius disc. End-of-year snapshots from 2013 to 2025 were obtained from the ohsome API, and income groups from the World Bank. City-level changes were tested with Wilcoxon signed-rank tests, bootstrap confidence intervals and Holm adjustment, and income gradients with Kruskal-Wallis tests and Spearman correlations. Results: Mapped parks per 100 km2 rose from a median of 31.7 to 135.6 (median change 99.6, 95% CI 77.7-226.6), and all 20 cities recorded an increase. Park area share, fitness stations, playgrounds and pitches also increased. Building polygons grew faster than parks, so parks per 10,000 buildings fell. Park area share differed by income group (p = 0.014), whereas park density did not (p = 0.084). Conclusions: Mapped park infrastructure expanded in every capital, but the expansion is inseparable from mapping effort and showed only a weak, imprecise association with income. Open map history can monitor the setting for walking-based augmented reality play if it is validated against official inventories.