A multi-dimensional urban spatial perception framework for urban diagnosis in Wuhan driven by multi-source data.

Area of Science:

• Urban Planning
• Geographic Information Systems (GIS)
• Computational Social Science

Background:

• Conventional urban design often overlooks lived experiences, focusing on functional aspects.
• Existing governance models lack comprehensive methods for evaluating multi-dimensional urban perception.

Purpose of the Study:

• To propose a multi-source data-driven framework for assessing urban spatial perception.
• To support urban diagnosis and grid-based governance by integrating environmental, social, and functional data.
• To address evaluation bias in urban planning and governance.

Main Methods:

• Integrating street-view imagery, mobile phone data, and POI semantics.
• Extracting environmental attributes and predicting affective perceptions (e.g., beauty, safety).
• Characterizing supply-agglomeration patterns and identifying functional rhythms.
• Constructing a "affective perception-agglomeration intensity-regional function" matrix.

Main Results:

• Classified Wuhan's urban space into 18 multi-dimensional perception units.
• Identified key areas requiring targeted, category-specific governance interventions.
• Developed a typology linked to governance-oriented references for different spatial types.

Conclusions:

• The framework provides a diagnostic basis for urban problem identification and intervention prioritization.
• It moves beyond functional provision to incorporate lived experience in urban spatial design.
• Offers a refined approach to urban governance and planning through multi-dimensional analysis.