Do peri-urban parks benefit low-income groups? Analyzing environmental equality with an integrated index and double machine learning

The growing public desire for interaction with natural environments has highlighted the importance of equitable access to peri-urban parks (PUPs) across diverse income groups. However, previous studies largely overlook the role of public perceptions in assessing PUP quality and fail to address endogeneity issues, leading to biased explorations regarding environmental equality. This study introduced a novel comprehensive index that integrates sentiment responses and visual preferences to reflect public perceptions. Using Chengdu as a case study, we applied a multi-mode Huff-based two-step floating catchment area model to evaluate the accessibility of residential communities to PUPs, with housing prices serving as a proxy for urban residents’ incomes. Additionally, a double machine learning approach was employed to estimate the treatment effect of housing prices on PUP accessibility, mitigating bias arising from endogeneity issues. The results reveal that (1) integrating social media text and image data comprehensively captures PUP quality. (2) Urban residents living outside the Outer Ring Road have better access to PUPs compared to those within the Outer Ring Road, with notable disparities in accessibility across the four cardinal directions. (3) A positive effect of housing prices on PUP accessibility is observed. Moreover, green gentrification occurs in certain regions, particularly in southern and eastern urban expansion zones, which often coincide with the focus points of urban development. These findings suggest that policymakers should enhance PUP equality by ensuring green spaces of affordable housing, allocating targeted funding for green space improvements in low-income areas, and enhancing transit access to underserved areas, thereby promoting a more equitable distribution of environmental benefits. © 2025 Elsevier GmbH