Using street view images to examine the association between human perceptions of locale and urban vitality in Shenzhen, China

There is a high correlation between the physical environment, human perception, and urban vitality. However, fine-scale variations in urban vitality are complex, and human perceptions of locale are difficult to measure. In this study, EasyGo data provided by Tencent, are used to distinguish differences in daytime and nighttime vitality in Shenzhen, China. Then, a series of subjective and objective variables is calculated to reflect human perceptions of locale based on street view images (SVIs). Finally, random forest and spatial lag regressions are adopted to analyze the driving forces of urban vitality. The results suggest that differences in urban vitality are manifestations of the unbalanced allocation of urban function, accessibility, building form, and human perceptions. The dominant variable category is urban function. There are obvious distinctions between daytime and nighttime vitality, particularly because the human perception category is increasingly important to nighttime vitality. This work sheds light on the relationships between human perceptions and urban vitality, providing suggestions for urban microrenewal and the construction of high-quality streets and liveable communities.