Vitality is a central metric for the successful renewal of waterfront industrial heritage, yet existing research often overlooks nighttime dynamics and the diverse needs of different demographic groups. This study proposes a multi-source quantitative workflow integrating street view imagery (SVI), online perception surveys, behavioral heatmaps, and explainable machine learning to evaluate both perceived and behavioral vitality. Focusing on regenerated industrial waterfront sites in central Guangzhou, the research identifies a distinct temporal shift in the drivers of urban life. Results show that daytime vitality is primarily influenced by greenness and street morphology, whereas nighttime vitality is driven by lighting conditions and social presence. Furthermore, the analysis reveals significant perceptual differences and specific environmental thresholds across various demographic groups. By linking measurable spatial attributes to time-sensitive and group-specific responses, this framework provides evidence-based support for the inclusive management of industrial heritage and nighttime economy planning. The study offers a reproducible method and transferable indicators for cross-city vitality assessments.
Huang et al. (2026) studied this question.