The impact of street-scale built environments on urban park visitations: A case study in Wuhan

The COVID-19 pandemic has changed human life globally. Existing studies have revealed that citizens' visitations to urban parks varied before and after the COVID-19 outbreak. However, few studies have examined how street-scale built environments (SBEs) on routes affect visitations to urban parks at varying COVID-19 risk levels. In this study, a stated-preference survey was conducted to investigate 3,218 visitors' changes in urban park visitation under various COVID-19 risk levels. In addition to park visit influencing factors, including park features, neighborhood built environment, socio-demographic attributes, and travel distances, multiple SBE indexes on visitors' routes to parks were obtained from 34,780 Baidu Map street view images using a deep neural network (DeepLabv3+) method. The results suggest that a high GVI and high traffic congestion on the route from the visitor's home to the urban park led to an increased probability of visiting the urban park by 188.1% (p = 0.044, OR = 2.881) and a decreased probability by 32.3% (p = 0.049, OR = 0.677), respectively. The high probability of visitation was also associated with socio-demographic attributes (including male gender, high income, high and medium education levels, and the elderly) and short travel distances.