Do different types of urban streets lead to varying COVID-19 risk perceptions? An empirical study from a spatial heterogeneity perspective

Abstract

People's risk perception varies across different environments. However, how different environments influence people's COVID-19 risk perception (RP), especially street environments, has yet to be fully evaluated. This study explores how three different types of street environments influence people's COVID-19 RP. First, we used pedestrian-centric street-view images and machine-learning models to predict the severity of COVID-19 RP in the study area. In addition, geographically weighted regression models were constructed to assess the locally most important environmental factors and their spatial heterogeneity. The results showed that openness was the most important influencing factor on COVID-19 RP in all three types of streets, while visual crowdedness was the least important one. Moreover, the high-value clusters of COVID-19 RP mostly occurred in poor- and medium-quality commercial streets, poor-quality residential streets, and leisure streets close to the plaza. Contrastingly, the low-value aggregations of COVID-19 RP mainly appeared in high-quality commercial streets, medium- and high-quality residential streets, and leisure streets close to blue space. We also found that the importance of environmental factors varied greatly by location. Interestingly, the RP was lower with higher visual crowdedness for all types of commercial streets. Based on this, the COVID-19 RP in different types of street environments was comprehensively evaluated so that city-relevant departments can take more effective measures to manage the public health risk based on geographical features.