A Smart Card Data-Driven Framework for Identifying Group Travel Patterns in Metro Systems

Abstract

Understanding how people travel together, rather than individually, is crucial for interpreting collective mobility behaviors in urban environments. This study proposes a data-driven framework to identify and characterize Group Travel Patterns (GTPs) in metro systems using large-scale smart card data. Unlike previous research that primarily focuses on individual travel behavior, this study highlights social synchronization and coordinated decision-making among passengers, thereby bridging behavioral analysis with transit operations. We identify two types of GTPs: Type A, representing occasional and socially motivated companionship, and Type B, reflecting regular, commuting-oriented joint travel. Based on one week of Beijing Metro smart card data, we systematically compare their temporal regularities, spatial clustering, and passenger attributes. Results show that Type B groups exhibit stronger temporal consistency and spatial concentration around residential and employment centers, whereas Type A groups demonstrate greater flexibility and leisure-oriented dispersion. These findings provide empirical evidence of collective behavioral regularities in public transport and offer practical insights for optimizing crowd management, fare policies, and service scheduling. Methodologically, this study contributes by combining proxemics-informed temporal thresholds with recurrence-based validation to distinguish intentional co-travel from coincidental encounters, and by validating the typology using cross-year AFC data—enhancements that improve detection precision relative to prior proximity-based methods.