Brief Introduction

Chinese urbanism has gradually shifted from growth-rate-oriented approaches to those prioritizing quality, disclosing many pressing issues within existing built environments, such as the lack of pedestrian-friendly street conditions. This research proposes a theoretical framework integrating both built environment assessment and pedestrian walking perception to assess street walkability in the historic urban landscape, as exemplified by the case study in Xiao'Qinhuai River historic area, Yangzhou. Embarking from the research gap drawn by the inaccuracy of vehicle-perspective streetscape on platforms, this study proposed a self-recorded database of streetscape imagery from the human perspective. A comprehensive methodological framework for the assessment was established then by leveraging the methods, including Space Syntax Analysis, Image Segmentation, Questionnaire Sample Survey, and Machine Learning. This research identifies and analyses the key factors influencing street walkability through the lens of street built environment and pedestrian walking perceptions. This framework introduces a replicable and transformable prototype for walkability assessment, creates an approach of data-driven, evidence-based decision-making in urban design, and highlights the crucial status of human-centric studies in urban regeneration.

Figure 1

Figure 2

Figure 3