All Times ET
Keywords: urban rail, congestion, causal statistical modelling, Bayesian machine learning, nonparametric statistics
Urban rails or metro systems are a major contributor to sustainable transportation in most cities. Metro systems often operate at high service frequencies to transport large volumes of passengers, specifically during peak hours. However, the reliability of such operations can be severely impacted by a vicious circle of high passenger demand and train delays. This paper proposes a novel model of the nature and form of the root cause of congestion delays in metro operations. The proposed approach leverages large-scale automated fare collection and train movement data to understand the metro's station-level congestion phenomenon. The method is transferable and provides a crucial tool to inform strategies for congestion management in metro systems.