From trajectories to network: Delineating the spatial pattern of recreational walking in Guangzhou

Recreational walking is increasingly recognized as an essential aspect of daily life, whose network structure serving as a crucial tool for characterizing activity patterns. However, existing methods lacks consideration of the heterogeneous spatial configuration when constructing recreational walking networks, thereby fall short in accurately identifying 'nodes' and establishing 'links'. This study aims to address these gaps by revising the network construction process. To do so, we incorporate the 'self-organized' nature of recreational activities and apply a 'resource' allocation method from user recommendation algorithms. Using downtown Guangzhou as a case study, the revised methodological framework can capture the striking spatial heterogeneity in walking activities. Such a heterogeneity is demonstrated by a pronounced rank-size effect in network 'nodes' and a notably low weighted clustering coefficient. Further, the detection of seven distinct communities outlines the broader spatial structure of these activities. The study concludes by exploring how this network analysis is capable of developing planning-based maps and enhancing ex-ante evaluations of urban planning initiatives.