“Love the Specific Scene of Life”: A Visual Analysis of User-Generated Narratives on Walkable City Spaces

Abstract

The increasing convergence of technology and media interactivity continues to shape novel urban landscapes and social connections, catalysing the emergence of place in media and the phenomenon of “Wang Hong urbanism”. This research investigates urban aesthetic appeal in the digital era through analysis of photographs posted on the social media platform Douban under #CityWalk, utilising a deep learning scene recognition model with Class Activation Mapping. The visual content analysis established three primary categories: natural landscapes, built structures, and human presence. This classification framework illuminates how distinct aesthetic elements are emphasised in user-generated urban representations. The findings demonstrate how pedestrians reconstruct their perceptions of walkable urban spaces through social media engagement, whilst providing insights into contemporary urban tourism within the context of Wang Hong’s urbanism.