Shared bicycles have emerged as a transformative force in urban transportation, effectively addressing the perennial 'last mile' challenge faced by commuters. The limitations of station-based bike-sharing systems, constrained by point-to-point travel, have spurred the popularity of the dockless model, offering flexible rentals and eliminating docking infrastructure constraints. However, the rapid growth of the sharing economy has introduced new challenges, notably an imbalance between supply and demand, leading to issues like the unavailability of bicycles and insufficient parking spaces during peak hours. To address these challenges, this study introduces a novel variable, Congestion Density (C), to quantitatively measure dynamic congestion levels in dockless bicycle-sharing systems. Leveraging real-time shared bike information from Xiamen, China, we present a sophisticated clustering framework for congested spots, identifying 563 congested spots categorized into Over-crowded, Semi-crowded, and Light-crowded clusters. Strikingly, these clusters align with established subway lines and bus stops, revealing a prevalent trend of integration between subway/bus services and bike-sharing. Overall, this study proposes parking lot management plans and policy recommendations based on the dynamics of crowded parking spaces, geographical characteristics, and land functional attributes. Our findings provide crucial insights for implementing bike-sharing electric fences and understanding urban mobility patterns, contributing to sustainable urban transportation.

翻译：共享单车已成为城市交通中的变革性力量，有效解决了通勤者面临的长期“最后一公里”难题。基于站点式共享单车系统受点对点出行限制的局限性，促使无桩模式日益流行，其提供灵活租赁并消除了停靠基础设施的约束。然而，共享经济的快速发展带来了新挑战，尤其是供需失衡，导致高峰时段车辆不可用及停车位不足等问题。为应对这些挑战，本研究引入新变量——拥堵密度（C），用以定量测量无桩共享单车系统中的动态拥堵水平。利用中国厦门的实时共享单车信息，我们提出了一个精细化的拥堵点聚类框架，识别出563个拥堵点，并将其分为过度拥挤、半拥挤和轻度拥挤三类聚类。引人注目的是，这些聚类与现有地铁线路和公交站点相吻合，揭示了地铁/公交服务与共享单车整合的普遍趋势。总体而言，本研究基于拥挤停车位的动态变化、地理特征和土地功能属性，提出了停车场管理计划与政策建议。我们的发现为实施共享单车电子围栏及理解城市出行模式提供了关键见解，有助于推动可持续城市交通发展。