Exploration of Urban Heritage Preferences in Chinese Context Using Computer Vision: An Analytic of Kulangsu International Settlement

Social media plays a core role in inclusive heritage management and contains multi-potential values to structure the inner connection of stakeholders and express the preferences of users and non-users in urban landscape studies without distinction. However, engaging residents to cultural heritage sites and understanding their preferences through social media have not been well-evaluated and explored in the Chinese context. The article offers a method to understand the residents’ preferences by identifying more than two thousand urban-heritage- linked images acquired from the dataset of Weibo. Both Google Cloud Vision detection and manual examination were utilized parallelly to ensure the validity of the result. Little difference was found in the comparison of results between the two judgement methods. The result revealed that residents in Kulangsu have a stronger interest in and concern more about the buildings and nature landscapes other than urban design. The study concludes with two points. First, the analysis of social media data is strongly recommended to be introduced in the decision-making process of urban heritage conservation as a strategy in the post-pandemic period. Second, computer vision can be a trustable tool to present residents’ preferences and is worth being widely applied in Chinese urban heritage studies.