Assessing the customer impact of service disruptions on the London Underground network using Automated Fare Collection data

Abstract

Service disruptions are a common undesirable occurrence in urban public transport networks, in response to which passengers often take action. Tis may involve changing their route, altering their origin and/or destination, switching to other modes or even cancelling their trip altogether. The aim of this study is to provide an insight into the factors that influence this behaviour. Using the London Underground network as an example, passenger responses to incidents are inferred by analysing an eight-week dataset of the “Oyster” Automated Fare Collection system, while service disruptions are extracted from London Underground’s CuPID database of incidents during the same period. Binary logistic regression is used to fit models describing passenger responses to disruptions in terms of continuing their journey, changing origin or destination station, or leaving the network altogether. The results suggest that passengers are more likely to take action in response to a service disruption if this has a delay of less than 5 mins or more than 20 mins, but more likely to stick to their original route for delay durations in between. Also, passengers are more likely to change station or leave the network if the disruption occurs at the origin station of their journey.