轨道交通站点周边土地利用与共享单车的耦合关系分析——以西安市为例

Analysis of the Relationship Between Land Use and Dockless Bikeshare Around Rail Stations——A Case Study of Xi’an

为促进共享单车接驳轨道交通出行和公交主导型的发展,利用GPS数据确定共享单车接驳轨道交通的范围,在该范围内基于POI数据,利用K均值算法将站点划分为四类。通过BP神经网络的方法研究四类站点下土地利用差异性和共享单车间的关系,并建立四个需求预测模型。结果表明在考虑不同轨道站点周边用地差异性的前提下,四个模型对共享单车需求的预测效果良好,说明轨道交通站点周边共享单车的配置需考虑站点类型和用地差异性。此外,研究结果发现大部分轨道站点对单车的吸引是来源于公司企业、商业服务、零售和住宅用地,且主要分布在市中心;单车数量分布较低的站点主要集中在首末站及其附近站,以及用地单一的站点。

In order to promote dockless bikeshares to connect rail transit, GPS data was used to determine the catchment areas. Based on POI data within this range, the K-means algorithm was used to divide the sites into four categories. A demand forecasting model based on BP neural network was established to explore the impact of land use on dockless bikeshares under four types of sites. The results show that the four models were effective in predicting the demand for dockless bikeshares under the premise of considering the difference in land use around the rail station. Explain that the configuration of dockless bikeshares around rail stations need to consider the differences in station types. In addition, the results of the study found that most of the track sites’ attraction to bicycles came from companies, commercial services, retail and residential land, and they were mainly located in the city center. The stations with a low number of bicycles were mainly concentrated at the first and last stations and nearby stations, as well as stations with a single land.