The Limits of AI-Driven Mobility Data in Children’s Leisure Research

Mobility data platforms such as Placer.ai and SafeGraph have gained rapid traction in park and recreation research. By aggregating anonymized smartphone location data, these AI-driven tools offer estimates of park visitation, visitor origin, and catchment areas that traditional methods, including SOPARC and intercept surveys, cannot capture. However, experience applying mobility data to park and playground visitation across U.S. cities reveals limitations that have yet to be addressed, particularly in children’s leisure research. The fundamental limitation concerns panel coverage and data representativeness: children are largely absent from opt-in smartphone panels, and low-income families are often underrepresented due to lower smartphone ownership rates. Mobility data perform reasonably well for large parks, but visitation data for smaller neighborhood playgrounds are often suppressed due to privacy thresholds. Finally, mobility data record presence, not access; transportation, safety, and park quality remain invisible. This commentary assesses these limitations and offers guidance for responsible use and reporting.