Examining streetscape visuals and emotional responses through social media and street view image analysis

Streets are important spaces for everyday activities, and the street environment can impact quality of life. This study investigates the impact of streetscape visuals on public emotions in the Beishan Street and Nanshan Road historic districts surrounding West Lake in Hangzhou.

Emotional perceptions were analyzed through a textual study of relevant posts on Sina Weibo, categorizing them into positive or negative sentiments. Simultaneously, Baidu Map Street View Images (SVIs) of the area were processed using the DeepLabV3 + Network model to semantically segment them into 19 elements, followed by the calculation of seven visual space indicators.

Correlation analysis between these street view elements, visual space indicators, and the emotional content of the posts revealed that multiple factors within the SVIs are significantly associated with public emotions. For instance, elements such as Rider and visual space indicators like Enclosure are positively correlated with emotions, while elements like Traffic Light and visual space indicators such as Openness show negative correlations. These emotional impacts vary depending on the specific building types present in the streetscape.

The findings underscore the association between streetscape visuals and public emotions and demonstrate that social media platforms provide substantial data for studying these effects. This research offers valuable insights for urban planners and managers to understand resident and tourist preferences, providing an objective foundation for enhancing urban streetscape quality.