Crafting a walkable and ridable urban green space system: Human-centered park network planning supported by active leisure travel trajectories

Urban green space (UGS) networks are essential for creating sustainable and livable cities, yet their planning agendas predominantly rely on ecological methodology, overlooking the behavioral connectivity of how people actively use and connect with these spaces. This study addresses this gap by developing a human-centered framework to analyze and inform UGS network design based on active leisure travel (ALT) patterns. Using a large dataset of GPS trajectories from Shenzhen, China, we construct and compare distinct park networks for walking and cycling based on co-visitation patterns. We apply network science methods, including node analysis and community detection, to reveal the structural properties and hierarchies of these behaviorally-defined networks. Our results reveal that walking and cycling form distinct network structures. Walking networks exhibit higher local clustering, while cycling networks are characterized by greater spatial distribution. The community-detection findings demonstrate the degree of alignment between top-down planning clusters and bottom-up travel network communities. Hierarchical linear models are further calibrated to examine the association between network structure and park-level activity, underscoring the significant role of park connectivity planning. This research provides a data-driven methodology for planners to evaluate UGS connectivity from an active traveler perspective, identify latent opportunities for network enhancement, and design more effective and activity-supportive green infrastructure.