Understanding the role of urban block morphology in innovation vitality through explainable machine learning

Innovative activities are a key driver of economic and social development, with urban blocks serving as essential hubs for innovation. However, how urban block morphology shapes innovation vitality remains challenging. This study uses spatial analysis and SHAP-based explainable machine learning to analyze how block morphology characteristics affect urban innovation vitality, taking Hangzhou, a typical city of digital innovation, as a sample. The findings show that: 1. Land use diversity, floor area ratio, and street network density are the most influential factors; 2. Block morphology exhibits nonlinear effects. For example, when street network density exceeds 200 m per square kilometer, its impact on innovation vitality is positive, but it diminishes beyond 400 m per square kilometer; 3. Synergistic effects are prevalent, such as when street network density exceeds 200 m per square kilometer and floor area ratio is between 0 and 7, enhancing innovation vitality. The results reveal the nonlinear effects and interaction mechanisms at the block scale. And also discusses potential planning strategies for different regions based on varying influencing factors. The study provides an effective approach to balancing precision and interpretability in spatial analysis and offers empirical support for the application of complexity science in urban studies.