Social media has enabled individuals to grasp the fragmented characteristics of urban places through photos and videos before physically exploring them, generating new patterns of urban imagination and environmental experience verification. On Chinese social media platforms, posts tagged with #citywalk number in the hundreds of thousands, reflecting the widespread appeal of this emerging leisure trend. To investigate this phenomenon, 203 participants used a Public Participation Geographic Information System (PPGIS) to map their walking routes and perceived landscape values within Beijing’s old town. Kernel density analysis identified five major City Walk hotspots along culturally and visually rich corridors. A Random Forest model showed that residential and road densities and POI diversity were the strongest predictors of City Walk intensity. Spatial analysis revealed that recreational and aesthetic values were most frequently perceived, followed by spiritual value and sense of place, with educational institutions, historical buildings, cafés and parks and scenic areas emerging as important POIs associated with these values. OLS and Bayesian regression analyses further confirmed the robustness of these findings, while out-of-sample validation demonstrated consistent model performance across multiple dimensions. This study offers a novel contribution by integrating PPGIS mapping with machine-learning and Bayesian modelling to reveal how built-environment features and demographic characteristics jointly shape City Walk behaviours and landscape value perceptions in the social media era. The findings provide clear policy implications for urban planning, highlighting the need to enhance street connectivity, protect cultural heritage, and design inclusive public spaces that support diverse social groups and encourage meaningful urban engagement.
Xu et al. (2026) studied this question.