考虑宏微观建成环境的共享单车骑行影响因素分析

针对共享单车骑行影响因素分析中对微观建成环境因素探讨的不足,以北京市五环内及附近街道为研究对象,基于城市多源数据,从公共交通设施、土地利用、人口经济属性、空间设计及道路环境5方面提取宏微观建成环境因素,并构建全局回归模型(global regression, GR)、地理加权回归模型(geographically weighted regression, GWR)及多尺度地理加权回归模型(multi-scale geographically weighted regression, MGWR),对共享单车骑行密度进行分析。研究结果表明：MGWR模型有更好的回归效果,调整拟合优度R²高达0.91;共享单车骑行密度空间上呈现“高-高”及“低-低”聚类两极分化现象;地铁站点密度为局部影响因素,公交线路密度及隔离护栏出现率为区域影响因素,其他建成环境变量为全局尺度;骑行密度显著影响因素的回归系数,在空间上较为平稳,单一变量的空间影响关系均为正向或负向影响;微观建成环境中,绿视率、交通信号出现率为正向影响,色彩丰富性及路灯出现率为负向影响。

Analysis of Factors Influencing Dockless Bike-sharing Cycling Considering Macroscale and Microscale Build Environment

Aiming at the insufficient discussion of microscale built environment in the analysis of dockless bike-sharing (DBS) cycling influence factors, macroscale and microscale built environment factors were extracted from the five aspects: public transportation facilities, land-use, socioeconomic attributes, spatial design and road environment in the sub-districts within and near the Fifth Ring Road of Beijing, based on urban multi-source data. Then, the global regression model (GR), geographically weighted regression model (GWR) and multi-scale geographically weighted regression model (MGWR) were constructed to analyze the DBS cycling density. The results show that the MGWR model has a better regression effect, and the goodness-of-fit adjusted R2 is as high as 0.91. The DBS cycling density obviously shows spatial differentiation of "high-high" and "low-low" clusters. The density of metro station is a local influence, the density of bus lines and the presence of barrier are regional influences, and other built environment variables are global scales. The significant influence coefficient of DBS cycling density is relatively smooth in space and the spatial influence relationships of single variables are all positive or negative. The greenery and the presence of signal in the microscale built environment factors positively affect DBS cycling density, while colorfulness and the presence of streetlight have a negative effect.