Exploring public experiences in urban green spaces: A social media analysis of Amsterdam

Urban green spaces (UGSs) are increasingly recognised for their ecological, social, and psychological benefits, particularly within rapidly urbanising contexts. Understanding public perceptions and interactions with UGSs is critical for realising these benefits and advancing sustainable urban development. This study employs crowd-sourced social media data, integrating textual and photographic reviews, alongside advanced machine learning techniques, to investigate public experiences in intra- and peri-urban green spaces in Amsterdam. Using a lexicon developed from cultural ecosystem services and landscape elements, the analysis identifies key characteristics and experiential differences between these two types of UGS. Results show that intra-UGSs are primarily valued for recreational infrastructure and social vitality, functioning as accessible community refuges, while peri-UGSs are appreciated for ecological quality, wildlife encounters, tranquillity, and nature-oriented experiences. Cross-modal comparisons reveal complementary dimensions. Textual reviews highlight affective impressions and functional use, while photographic reviews capture visually prominent features. Cluster analyses further classify the 32 UGSs into five categories: balanced, animal-themed, vegetation-themed, recreation-oriented, and architecture-dominated. This reveals perceptual distinctions beyond conventional location- or function-based typologies. This novel classification exposes the hidden diversity of UGSs and provides methodological innovation and actionable insights for adaptive, user-centred planning and management.