Unveiling the Spatial Mismatch Between Green Space Equity and Residents’ Subjective Well-Being: An Integrated Approach Based on Machine Learning and Social Media Data

The limited capacity of urban green spaces to equitably satisfy the well-being needs of populations in urbanized areas is a global challenge. However, research on the spatial mismatch between green space equity and residents’ subjective well-being (SWB) remains inadequate. Using Shanghai as a case study, this research integrates social media data with an improved GA2SFCA method to evaluate SWB and UPGS accessibility and analyzes and compares the geographical spatial distribution differences of UPGS accessibility across different travel modes. This study employs machine learning to reveal the potential drivers of the mismatch between SWB and UPGS accessibility (note that this study does not explore causal relationships). The results indicate that: (1) UPGS accessibility in Shanghai exhibits pronounced spatial heterogeneity, the equity results derived from the Lorenz curve and Gini coefficient indicate that public transit (Gini = 0.579) < walking (0.427) < driving (0.149), and community parks effectively mitigating disparities among other urban park types; (2) UPGS accessibility and SWB are spatially correlated (r = 0.013, p < 0.01, z > 2.58), with a distinct High-High clustering pattern identified in the inner-ring region; (3) Road network accessibility (SHAP = 0.9478), housing prices (0.7025), and company agglomeration (0.5695) are the three most influential factors contributing to the spatial mismatch where SWB is higher than accessibility, and they exhibit clear threshold effects. These findings link urban green space equity with residents’ SWB, providing a basis for targeted interventions to enhance social welfare and promote urban sustainability.