Designing public transit services that meet the needs of night shift workers requires understanding their commuting patterns. However, traditional survey methods have faced challenges in contacting and interviewing night shift workers with changing work–sleep schedules. This study aims to investigate night shift workers’ commuting patterns by identifying night shift workers with heterogeneous working patterns in passive mobility data. First, we identify workers and their workplaces from the mobility data using a rule-based method. Subsequently, we cluster individual workers’ workplace-staying records using DBSCAN to find regular working patterns and segment workers with diverse working patterns. We applied the method to the smart card data of Seoul, South Korea, and identified 37,448 night shift workers with six different working patterns. The proportion of night shift workers among presumed workers was 8.9%, which was slightly higher than the 7.2% reported in a national survey. Workers who exclusively worked at nighttime worked near night markets, while three-shift workers’ workplaces mainly appeared near areas with large general hospitals with emergency care centers. The proportion, regular working patterns, and workplace location distribution of the identified night shift workers were generally consistent with the existing survey-based knowledge, suggesting that night shift workers were accurately discovered from the mobility data. The finding suggested that policies to extend transit service time were beneficial for three-shift and 24-hour duty workers. With the proposed method, night shift workers’ regular and essential mobility needs can be identified and provided to cities’ transit authorities to help assess and improve transit services for these workers.

中文翻译：

---

使用被动出行数据调查夜班工人的通勤模式

设计满足夜班工人需求的公共交通服务需要了解他们的通勤模式。然而，传统的调查方法在联系和采访工作睡眠时间不断变化的夜班工人时面临着挑战。本研究旨在通过识别被动出行数据中具有异构工作模式的夜班工人来调查夜班工人的通勤模式。首先，我们使用基于规则的方法从流动数据中识别工人及其工作场所。随后，我们使用 DBSCAN 对单个工人的工作场所停留记录进行聚类，以找到规律的工作模式，并对具有不同工作模式的工人进行细分。我们将该方法应用到韩国首尔的智能卡数据中，识别出 37,448 名夜班工人，他们有六种不同的工作模式。夜班工人占推定工人的比例为8.9%，略高于全国调查报告的7.2%。专门从事夜间工作的工人在夜市附近工作，而三班制工人的工作场所主要出现在设有急救中心的大型综合医院附近。所识别的夜班工人的比例、规律工作模式、工作地点分布与现有的调查知识基本一致，表明从流动数据中准确地发现了夜班工人。研究结果表明，延长公交服务时间的政策对三班制和 24 小时值班工人有利。通过所提出的方法，可以确定夜班工人的常规和基本出行需求，并将其提供给城市交通当局，以帮助评估和改善这些工人的交通服务。