基于TOD模式及语义功能的共享单车潮汐流与建成环境非线性关联研究

Nonlinear Correlation between Shared Bike Tidal Flow and Built Environment under the Guidance of Transit-Oriented Development and Semantic Function

公共交通导向发展(TOD)是实现城市紧凑、人居可持续发展的重要途径,共享单车与轨道交通的接驳出行极大提升了公共交通的服务范围,促进了TOD规划策略的落实。该文基于TOD规划理念,识别共享单车的接驳流动模式,在考虑地铁站语义功能基础上,使用可解释机器学习模型挖掘武汉市主城区共享单车潮汐流模式及与建成环境的非线性关联关系,结论如下:①根据共享单车早晚高峰时段的潮汐均衡性指数,发现武汉市主城区存在“早入晚出”和“早出晚入”两类潮汐流模式,前者出现在距城市中心距离中等的居住区和汉口中央活动区边缘的就业区,后者出现在二环线附近的就业密集区;②通过POI语义信息建模和层次聚类识别出4类地铁站语义功能主题及两种混合功能聚类簇,地铁站语义功能类型与共享单车流动模式呈强相关;③建成环境因子的非线性效应明显,当配套生活设施和道路密度适中、居住规模较大时,能有效促进共享单车高流量稳定接驳模式发生。研究结果可从共享单车优化调度、城市建成环境调整等方面为缓解共享单车潮汐流问题提供依据。

Transit-oriented development(TOD)is a crucial way to achieve compact,livable,and sustainable urban development.The integration of shared bikes and rail transit significantly expands the service scope of public transportation and promotes the implementation of TOD planning strategies.Based on TOD planning concepts,this paper identifies the transfer flow patterns of shared bikes.Considering the semantic functions of metro stations,an interpretable machine learning model is used to explore the tidal flow patterns of shared bikes and their nonlinear correlations with the built environment in the main urban area of Wuhan.The conclusions are as follows.①Based on the tidal balance index of shared bikes during morning and evening peak periods,two types of tidal flow patterns are identified in the main urban area of Wuhan,namely"convergence in the morning and divergence in the evening"and"divergence in the morning and convergence in the evening".The"convergence in the morning and divergence in the evening"tidal pattern mainly occurs in residential areas at a moderate distance from the urban center and employment areas on the edge of the Hankou central activity zone,while the"divergence in the morning and convergence in the evening"tidal pattern mainly occurs in employment-dense areas near the Second Ring Road.②Four types of semantic functional themes and two mixed functional clusters of metro stations are identified through POI semantic information modeling and hierarchical clustering.The semantic functional types of metro stations are strongly correlated with the flow patterns of shared bikes.③The nonlinear effects of built environment factors are significant.The moderate supporting living facilities and road density,and the large residential scale can effectively promote the occurrence of stable connection pattern of shared bikes with high riding flow.This study can provide a basis for alleviating the tidal flow problem of shared bikes from the aspects of optimizing shared bikes scheduling and adjusting the urban built environment.