Urban transportation and planning are at a pivotal juncture, requiring a nuanced understanding of the built environment's impact on dockless bike sharing (DBS) to metro transfer trips. Existing methodologies, often focused on DBS trips around metro stations or limited by scant datasets, overlook the pressing need for a method to discern large-scale DBS-metro trips as DBS becomes a standard feeder mode for metro trips and the yet unexplored spatial lag effects of the built environment on DBS-metro interactions. To bridge these gaps, we develop a method integrating smart card records and DBS data, revealing a comprehensive trip chain encompassing both metro and DBS. Our application of association rule algorithms to large-scale data provides detailed spatial insights into feeder trips. We employ a network-adjacency-based partial spatial Durbin model, tailored with a negative binomial regression for count data and maximum likelihood estimation for continuous data. Analysis from Shenzhen reveals: (1) A strong correlation is observed between the count of trips using cycling as a feeder mode (COUNT) and the location of stations within the metro network structure. Notably, the COUNT shows more significant aggregation when compared to the ratio of DBS-metro transfer trips to the total metro trips at each station (RATIO); (2) significant influence of both local and adjacent spatial variables of the built environment on the RATIO and COUNT of cycling trips; (3) specific factors like feeder station location, city center proximity, Street greenness view situation, and road intersection density significantly influencing the cycling feeder mode for metro trips; (4) Moreover, areas with more urban villages and industry appeared to contribute to the cycling feeder mode for metro trips, both in terms of RATIO and COUNT. This study underscores the necessity of fostering a conducive built environment to leverage DBS's potential to bridge the last-mile gap.

中文翻译：

---

整合智能卡记录和无桩自行车共享数据，了解建筑环境对自行车作为地铁出行支线模式的影响


城市交通和规划正处于关键时刻，需要细致入微地了解建筑环境对无桩共享单车 (DBS) 到地铁换乘出行的影响。现有的方法通常侧重于地铁站周围的 DBS 出行或受到数据集不足的限制，忽视了随着 DBS 成为地铁出行的标准支线模式以及尚未探索的空间滞后效应，迫切需要一种识别大规模 DBS 地铁出行的方法星展银行与地铁互动的建筑环境。为了弥补这些差距，我们开发了一种集成智能卡记录和星展银行数据的方法，揭示了涵盖地铁和星展银行的全面出行链。我们将关联规则算法应用于大规模数据，提供了对支线行程的详细空间洞察。我们采用基于网络邻接的部分空间 Durbin 模型，该模型针对计数数据进行负二项式回归，并针对连续数据进行最大似然估计。深圳的分析表明：（1）以自行车为支线方式的出行次数（COUNT）与地铁网络结构中车站的位置之间存在很强的相关性。 值得注意的是，与星展银行地铁换乘出行与每个车站地铁总出行的比率（RATIO）相比，COUNT 显示出更显着的聚合； (2) 建成环境的局部和邻近空间变量对骑行出行的RATIO和COUNT有显着影响； (3) 接驳站位置、距离市中心的远近、街道绿化景观情况、道路交叉口密度等具体因素对地铁出行的自行车接驳模式有显着影响； （4）此外，无论是在RATIO还是COUNT上，城中村和工业较多的地区似乎对地铁出行的自行车接驳模式做出了贡献。这项研究强调了培育有利的建筑环境的必要性，以利用星展银行的潜力来弥合最后一英里的差距。