The discrepancy between the physical accessibility and perceived accessibility of public service facilities in coastal towns and villages of Fujian, China, was investigated in this study. Through a spatial distribution analysis of education and medical service facilities in Liushui Town and Xingchen Town, the equity and coverage levels of various facilities were evaluated based on the residents’ physical travel costs and perceptions. The results show pronounced spatial heterogeneity between the physical and perceived accessibility across different regions. Our findings suggest that, while certain areas boast a lot of physical access to facilities, the residents’ perceived access is significantly influenced by factors such as the terrain, transportation conditions, and the types of available travel tools. The findings of this study provide a scientific basis for optimizing the allocation of rural public service facilities, aiming to bridge the foundational service gap between urban and rural areas and promote the equitable development of rural living environments.

Urban noise complaints reflect not only perceived acoustic disturbance, but also complaint behaviour shaped by thermal conditions and the built environment. Using Sanya, China, as a case study, this study integrated Landsat-derived land surface temperature data with the spatial distribution of noise complaints to examine how thermal environment and urban contextual factors jointly influence complaint patterns. An interpretable modeling framework combining eXtreme Gradient Boosting (XGBoost) and Multiscale Geographically Weighted Regression (MGWR) was employed to assess the associations of urban heat island intensity (UHI), road density, point of interest (POI) count, and population density on complaint occurrence and intensity, while also generating spatial predictions of complaint distribution. The results revealed a weak but statistically significant spatial association between the thermal environment and noise complaints, with urban heat island intensity showing nonlinear and spatially heterogeneous associations with complaint counts. POI count emerged as the strongest global predictor, while road density, population density, and thermal environment exhibited substantial spatial heterogeneity in their associations with complaint patterns. The integrated model outperformed both individual models alone, achieving the highest prediction accuracy. Overall, the findings suggest that urban noise complaint patterns reflect not only perceived acoustic disturbance, but also context-dependent social perception and reporting behaviour, providing empirical support for more place-sensitive and people-centered urban noise governance.