Ecological well-being performance in a Chinese urban agglomeration: Spatiotemporal analysis and policy insights from an Orange-based machine learning framework

From the global perspective of human–nature sustainable development, ecological well-being performance (EWP) captures the efficiency dimension of sustainability by linking ecological consumption with human well-being. For a rapidly developing country, how to balance development and ecology in the spatial dimension is not only related to spatial coordination and sound policy-making, but also determines the sustainability of growth. However, existing studies often overlook spatial disparities, typological differentiation, and nonlinear determinants. To address these gaps, this study developed a machine learning-based multi-method framework on the Orange visual programming platform, integrating spatiotemporal analysis and interpretable machine learning, with a focus on a typical region of China as the study object. This framework was applied to examine EWP in the middle reaches of the Yangtze River urban agglomeration (MRYRUA) during 2005–2022. The results show an overall improvement but with persistent spatial disparities, notable typological differences among cities, and key drivers dominated by industrial structure and government expenditure. By revealing the disparities behind aggregate progress, this study contributes to more precise estimation results and a clearer understanding of driving mechanisms, while accurately restoring spatial patterns, thereby laying a solid foundation for scientifically formulating multi-scale regional development policies.