地理科学进展

地理科学进展 ›› 2019, Vol. 38 ›› Issue (6): 791-806.doi: 10.18306/dlkxjz.2019.06.001

• 研究论文 •    下一篇

城市轨道交通通勤与职住平衡状况的关系研究——基于大数据方法的北京实证分析

申犁帆1(), 张纯2,*(), 李赫3, 王烨4, 王子甲5   

  1. 1. 武汉大学城市设计学院,武汉430072
    2. 北京交通大学建筑与艺术学院,北京 100044
    3. 中国银行国际金融研究所,北京100818
    4. 广州市城市规划勘测设计研究院,广州 510030
    5. 北京交通大学土木建筑工程学院,北京 100044
  • 收稿日期:2018-07-10 修回日期:2018-12-10 出版日期:2019-06-28 发布日期:2019-06-27
  • 通讯作者: 张纯
  • 作者简介:第一作者简介：申犁帆(1987—),男,北京人,博士,研究方向为城市轨道交通与空间规划、城市居民职住关系、城市可持续发展。E-mail: <email>495062785@qq.com</email>
  • 基金资助:
    国家自然科学基金项目(51678029,51778039);中国城市轨道交通协会专项研究项目(A17M00080)

Relationship between urban rail transit commuting and jobs-housing balance: An empirical analysis from Beijing based on big data methods

Lifan SHEN1(), Chun ZHANG2,*(), He LI3, Ye WANG4, Zijia WANG5   

  1. 1. School of Urban Design, Wuhan University, Wuhan 430072, China
    2. School of Architecture and Design, Beijing Jiaotong University, Beijing 100044, China
    3. International Finance Institute, Bank of China, Beijing 100818, China
    4. Guangzhou Planning & Design Survey Research Institute, Guangzhou 510030, China;
    5. School of Civil Engineering, Beijing Jiaotong University, Beijing 100044, China
  • Received:2018-07-10 Revised:2018-12-10 Online:2019-06-28 Published:2019-06-27
  • Contact: Chun ZHANG
  • Supported by:
    National Natural Science Foundation of China, No. 51678029 and 51778039;Specific Research Project of China Urban Rail Transit Association, No. A17M00080.

RichHTML

PDF (PC)

赞

摘要：

城市轨道交通网络的发展在提高居民通勤效率的同时也对其职住平衡状况产生了一定影响。论文以北京市206个轨道站点为例,基于高斯混合模型(Gaussian mixture model, GMM)和一卡通刷卡数据将轨道站点按职住功能进行分类,利用腾讯“宜出行”定位数据考察轨道站点周边的动态人口分布并计算就业居住比。研究发现：① 中心城区的职住状况明显优于中心城区以外区域;② 轨道交通线网末端区域的职住平衡程度较差,仅有少数成规模的高端服务产业集中分布的轨道站点周边形成了区域性就业中心;③ 部分就业-居住较为均衡的城郊地区仍存在一定的职住不匹配现象。随后,通过计算一卡通出进站比和“宜出行”职住比得到出进站均衡度和职住平衡度,利用广义自回归条件异方差(generalized autoregressive conditional heteroskedasticity, GARCH)模型对轨道交通通勤和职住平衡程度进行相关性分析,研究结果表明：① 出进站均衡度与职住平衡度具有非常显著的正向关系,即站点进出站人数越接近,站点周边区域的职住状况越好;② 典型就业地站点与站点周边区域的职住平衡程度显著正相关,而典型居住地站点与站点周边区域的职住状况存在显著的负相关性。这表明,人口稠密的聚居区无法带动同样数量就业岗位的产生,而完善的就业中心能够吸引一定数量的人口在附近居住;③ 轨道站点的区位条件与职住平衡状况存在一定正向关系;④ GMM能够对属性复杂模糊的轨道站点进行有效的聚类分析;⑤ 具有实时性强、精确度高、覆盖度广、获取难度低等优点的“宜出行”数据能够在微观空间尺度下弥补其他捕捉和分析实时人口时空分布特征方法的局限性。

关键词: 城市轨道交通, 通勤行为, 职住平衡, 大数据, 高斯混合模型, GARCH模型, 北京

Abstract:

The development of urban rail transit (URT) network improves the commuting efficiency of residents while it has a certain impact on their jobs-housing balance. This study took 206 URT stations in Beijing as an example and classified them according to their jobs-housing functions based on the Gaussian mixture model (GMM) and smart card data. The dynamic population distribution characteristics around URT station were explored and jobs-housing ratio was calculated by "Yichuxing" position data. The study found that: 1) The jobs-housing balance in the central city is obviously better than that outside of the central city. 2) At the ends of the URT network, the jobs-housing balance is worse while only a few stations with concentrated distribution of top service industries have formed regional employment centers. 3) There still exists a certain degree of jobs-housing mismatch in the areas around some suburban stations where employment and residential functions are relatively equal. Station outflow-inflow and jobs-housing balances were calculated by the station egrass-ingrass ratio and the jobs-housing ratio, and the correlation between URT commuting behavior and jobs-housing balance was analyzed by generalized autoregressive conditional heteroskedasticity (GARCH) model. The results of this study indicate that: 1) There is a very strong positive relationship between URT station egrass-ingrass balance and jobs-housing balance. The closer the numbers of URT station outflow and inflow population, the better the jobs-housing balance around the URT station is. 2) There is a strong positive relationship between employment opportunity and jobs-housing balance around a URT station; and there is a strong negative relationship between residential function and jobs-housing balance around a URT station. This suggests that dense settlement will not generate the same quantity of jobs while well-developed employment hubs can attract a certain number of residents to live nearby. 3) There is a positive correlation between locational conditions of URT stations and jobs-housing balance. 4) The GMM can effectively cluster URT stations with complex and unclear attributes. 5) With its advantages of real-time data capturing, high precision, wide coverage, and great accessibility, "Yichuxing" position data can effectively compensate for the limitations of other methods on collecting and analyzing spatial-temporal characteristics of real-time population distribution at the microscopic scale.

Key words: urban rail transit, commuting behaviour, jobs-housing balance, big data, Gaussian mixture model, generalized autoregressive conditional heteroskedasticity (GARCH) model, Beijing