A multi-scale user-friendliness evaluation approach on cycling network utilizing multi-source data

Cycling is a promising solution to transportation decarbonization and urban sustainability. To enhance the cycling environment, it's essential to assess the user-friendliness of cycling networks. Existing related studies lack large-scale quantitative assessments and typically focus on segments, which are cost-ineffective for extensive spatial coverage and disregard the comprehensive impact of the cycling network system. Therefore, we propose a multi-scale user-friendliness evaluation approach for cycling networks utilizing multi-source geospatial data. First, we construct the cycling network using bike-sharing trajectory data. Then, to assess the user-friendliness of individual cycling segments as transportation spaces, we derive and evaluate six fine-grained quantitative indicators of bike lanes (separation, shade, width, pavement, slope undulation, and connectivity) from streetview images, DEM, and the aforementioned cycling topological network. Thirdly, cycling communities, representing inhabitants' daily living spaces, are detected based on trajectory density that informs cycling patterns, and their user-friendliness is evaluated by combining the performance of cycling segments with the spatial distribution of POI facilities within each community. An empirical study in Xiamen, China demonstrates the effectiveness of the proposed method in pinpointing specific indicators at particular segments/underdeveloped communities, thus guiding investment and optimization of cycling infrastructure to promote greater adoption of cycling as a mode of transportation.