As extreme heat events become increasingly frequent worldwide, there is an urgent need for fine-scale assessment of urban heat risk and for identifying its key determinants. Conventional approaches often struggle to capture complex intra-urban spatial heterogeneity, limiting effective heat risk governance and resource allocation. This study applies the Hazard–Exposure–Vulnerability–Adaptation (HEVA) framework by integrating remote sensing, road network, and socio-demographic data. Using the CRITIC weighting method, we quantify and map a street-level heat risk index (HRI) in Tianjin, China. We further employ geographically weighted machine learning models to identify dominant drivers and to characterise nonlinear effects, interaction patterns, and spatially varying relationships. Model reliability is assessed by benchmarking geographically weighted models against global nonlinear baselines under three-fold cross-validation; GW-XGBoost achieves comparable explanatory power to the best global model (R² = 0.672) while yielding lower prediction errors (MAE = 0.142), supporting robust spatial inference. Results show that elevated heat risk is not confined to the urban core; instead, it is more pronounced in peripheral transitional zones around central districts. These areas often exhibit coincident heat stress and high population exposure, a higher concentration of vulnerable groups and ageing residential neighbourhoods, and comparatively limited access to medical and cooling resources. Mechanistically, greater development intensity is generally associated with higher heat risk, whereas higher vegetation cover tends to reduce risk; however, the strength and, in some locations, the direction of these effects vary substantially across streets. These findings suggest that heat risk management should prioritise peripheral transitional zones. Targeted interventions should balance development intensity, expand effective greening and shading, and improve the provision and accessibility of healthcare and cooling services to reduce street-level heat risk.

Image-based streetscape assessments commonly report separate perception scores or an aggregate index. Joint profiles additionally identify which unfavorable scores occur together. We examine joint lower-tail profiles of transferred safety, beauty, and liveliness estimates in 2304 grid cells represented by 48,011 street-view records from central urban Shanghai. A Perceptual Streetscape Assessment Chain separates visible content, estimated impressions, joint shortfalls, and contextual distribution. The Spatial Multidimensional Tail-Equity (SMTE) model represents eight profiles jointly and derives compound and dimension-specific probabilities with multilevel and spatial structure. At the primary lower-quartile threshold, 24.44% of cells had at least two low scores. Strong empirical overlap between safety and beauty made domain construction central to the interpretation of this classification. In this specification, the clearest conditional price-rank difference concerned low liveliness; its interval-based support varied across alternative thresholds. Mathematical exchanges toward greenery were associated with lower fitted compound risk under the specified composition model. The contribution is a profile-based screening approach that makes dependence and identification choices explicit. Its findings concern relative model-estimated visual conditions and housing-price contexts within the sampled central urban area, with resident ratings and household circumstances defining further targets for locally grounded evaluation.

Perceived street vitality directly reflects residents’ assessments of the attractiveness of the street environment; it is not only an important focus of urban vitality research but also closely related to human-centred sustainable urban development. However, limited data availability and the complexity of urban environments have constrained fine-grained spatial analysis at the city scale. To address this issue, this study quantified perceived street vitality by collecting street-view imagery, extracting streetscape features, and integrating these data with questionnaire survey results. After comparing multiple models, a geographically weighted machine learning model was employed to identify key visual predictors, model-estimated marginal associations, interaction patterns, and spatial heterogeneity related to perceived street vitality. The results show that areas with high perceived street vitality are mainly located along street segments with abundant greenery and open spaces, whereas low-value areas are concentrated in densely built and enclosed environments. Among the various streetscape elements, buildings, vegetation, and sky are the key visual elements most strongly associated with perceived street vitality. A model incorporating these elements accounted for 67.2% of the variance in perceived street vitality. Notably, the strength of these associations varied significantly across different areas. This study provides empirical evidence and evidence-based support for sustainable urban renewal, the optimisation of street-space layouts in high-density urban areas, and the improvement in street environmental quality.