Measuring Urban Green Space Vitality through Multi-Source Visual and Textual Data: Integrating Social Media and Street- Level Imagery in London

Abstract. Urban green space (UGS) vitality reflects actual public engagement rather than mere spatial provision. However, limited research has examined how micro-scale environmental characteristics shape participation intensity. This study investigates the relationship between street-level environmental features and green space vitality using cross-platform social media data in London. Geotagged Flickr and Instagram posts were used as proxies for usage intensity. Environmental attributes were extracted from Google Street View imagery through semantic segmentation, including vegetation coverage, walkability, spatial enclosure, and facility elements. Ordinary Least Squares models were applied to assess associations between environmental variables and social media density. Results indicate that structural and facility-related features show stronger relationships with participation intensity than simple vegetation measures.