Urban Green Spaces (UGS) are crucial to urban ecosystems, and understanding public perception is key to their effective management. While conventional survey methods are resource-intensive, Social Media Data (SMD) offers a cost-effective alternative to gathering public insights. However, current explorations into SMD's potential in assessing UGS face challenges, particularly in integrating text and image analysis and effectively filtering out irrelevant content to identify influencing factors of different dimensions. Based on a manually curated dataset, this study introduces the Park Dual-modal Perception (PDP) model, a cutting-edge approach combining SMD text and image analysis for evaluating perceived park accessibility, usability, and attractiveness, with an average accuracy of 86.81%. Utilizing SMD from 130 parks in Guangzhou, the model effectively quantifies the three dimensions, generating visualized scoring maps to aid planners in identifying parks with lower perceived scores on the urban scale. Further incorporation of SHapley Additive exPlanations (SHAP) within the PDP model can filter 82.79% irrelevant words and effectively extract 158 thematic words and 954 associated words, providing suggestions for park-level enhancement. Our findings indicate that (1) factors such as distance, travel time, ticket prices, and proximity to commercial amenities are pivotal in determining park accessibility. (2) Park usability hinges on park’s ability to serve diverse groups and provide well-maintained, multifunctional facilities. (3) Park attractiveness is closely linked with the cultural and regulatory characteristics of ecosystem services. Our methodology combines assessment and interpretation of human perception at both city and park scales. It aids city decision-makers in identifying low-quality parks and understanding the underlying reasons, thereby facilitating more informed urban planning decisions.
 

Urban green space,Public perception,Natural Language Processing,Computer vision,Social media data