空间异质性视角下常规公交客流影响因素

常规公交出行是城市客运体系重要的组成部分，其出行强度明显受城市建成环境的影响作用。以往建模方法未考虑影响因素之间的多尺度空间异质性问题，研究结果均存在一定局限性。以青岛市主城区为研究对象，首先基于公交GPS与IC卡数据提取常规公交客流为研究的因变量，并通过网络地图开放平台获取14个指标作为自变量；其次，通过探索性空间数据分析探究303个研究单元的常规公交客流空间演化特征；最后，分别构建地理加权回归与多尺度地理加权回归模型量化分析建成环境变量对常规公交客流影响的空间异质性。研究结果表明：与地理加权回归模型相比，多尺度地理加权回归模型的对回归结果的解释性更强；学校密度、医院密度、公交线路密度和停车场密度正向影响早晚高峰公交客流量，购物密度、土地利用混合度、距公交站点距离对早晚高峰的常规公交客流均有显著负向影响；公交线路密度、学校密度具有明显的空间异质性，对常规公交客流的影响程度整体上呈现由南向北扩散性递减的趋势。

Influencing Factors of Conventional Public Transport from the Perspective of Spatial Heterogeneity

Conventional public transport travel is an important part of the urban passenger transport system, and its travel intensity is obviously affected by the urban built environment. Previous modeling methods did not consider the multiscale spatial heterogeneity between the influencing factors, and the research results have certain limitations. Taking the main urban area of Qingdao as the research area, firstly, extract conventional public transport passenger flow from public transport GPS and IC card data as the dependent variables of the study, then obtain 14 indicators as independent variables through the open network map platform. Spatial data is used to analyze the spatial evolution characteristics of conventional public transport passenger flow of 303 research units. Finally, geographically weighted regression and multiscale geographically weighted regression model were introduced to quantitatively analyze the spatial heterogeneity of the impacts of built environmental variables on the passenger flow of conventional bus. The results show that the multiscale geographically weighted regression model is more explanatory to the geographically weighted regression model; the school density, hospital density, bus line network and parking lot density positively affect the bus traffic, shopping density, land use mixing and distance from bus station significantly negatively affect the conventional public transport passenger flow in morning and evening peak hours, and the bus line network and school density have obvious spatial heterogeneity and their influence degree shows a trend of diminishing spread from south to north.