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，以了解建筑环境对作为地铁出行馈线模式的骑行影响


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自行车共享 （DBS） 到地铁换乘行程的影响。现有的方法通常集中在地铁站周围的 DBS 行程或受到数据集不足的限制，随着 DBS 成为地铁出行的标准馈线模式以及尚未探索的空间滞后效应，人们忽视了对识别大规模 DBS-地铁出行的方法的迫切需求。为了弥合这些差距，我们开发了一种整合智能卡记录和 DBS 数据的方法，揭示了一个包括 Metro 和 DBS 在内的综合行程链。我们将关联规则算法应用于大规模数据，提供了对馈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，对计数数据采用负二项式回归，对连续数据采用最大似然估计。来自深圳的分析显示：（1） 使用自行车作为馈线模式 （COUNT） 的出行次数与地铁网络结构中的车站位置之间存在很强的相关性。 值得注意的是，与 DBS-metro 换乘行程与每个车站的地铁总行程之比 （RATIO） 相比，COUNT 显示出更显著的聚合;（2） 建筑环境的局部和相邻空间变量对自行车出行的 RATIO 和 COUNT 有显著影响;（3）支线站位置、市中心距离、街道绿化景观情况、道路交叉口密度等特定因素对地铁出行的自行车支线模式有显著影响;（4） 此外，城中村和工业较多的地区似乎有助于地铁出行的自行车馈线模式，无论是在 RATIO 还是 COUNT 方面。这项研究强调了营造有利建筑环境的必要性，以利用 DBS 的潜力来弥合最后一英里的差距。