1 Introduction

The COVID-19 pandemic created one of the largest disruptions to urban mobility in recent history. Lockdowns, restrictions on movement, and widespread adoption of remote working reshaped daily travel patterns in cities (Batty, 2023) across the world, with public transport systems among the most affected. These changes were not voluntary but imposed, creating conditions that scholars have described as “forced experiments” in travel behaviour. In transport research, a forced experiment refers to situations where external shocks force people to change their behaviour, creating a natural experiment without deliberate intervention (Larcom et al., 2017). The London Underground has previously been studied through this lens: Larcom and colleagues showed that a two-day strike left a measurable and lasting impact on commuting choices. If such a short disruption can alter travel behaviour, the pandemic, which persisted for months and years, raises the question of whether its impact on Underground ridership would be long-lasting, or whether demand would eventually revert to pre-existing trends.

The London Underground provides a particularly useful case study for examining this question. As the backbone of commuting in London, it carried around four million passenger journeys per day before the pandemic, and its financial model relies heavily on fare revenue. Disruptions to ridership therefore have significant implications not only for mobility, but also for the wider sustainability of the transport system. Central London stations, concentrated in Zone 1, are strongly associated with work and leisure activities, while peripheral stations in outer fare zones serve more local and residential trips. This central–peripheral divide makes the network an informative setting for assessing whether recovery has been geographically uneven. In addition, Transport for London (TfL) publishes annualised station entry and exit data, enabling consistent measurement of ridership across the entire system. By 2025, this dataset covers seven years, from 2017 through to 2023, spanning both pre-pandemic stability and post-pandemic recovery.

A growing body of research has investigated the impacts of the pandemic on London’s mobility. Trasberg and Cheshire (2021) highlighted the collapse of central activity hubs during lockdown, while Gao et al. (2023) found substitution between Underground trips and increased use of shared bicycles. Murcio et al. (2023) examined Origin–Destination commuting flows, showing that commuting remained below pre-pandemic levels well into 2022. Other studies have taken behavioural approaches, such as Bansal et al. (2021, 2022), who analysed risk perceptions and their effects on Underground ridership. While these studies have advanced understanding of pandemic mobility, few have made direct use of TfL’s annual station-level Entry and Exit data. This dataset provides a consistent, network-wide view of Underground demand, allowing recovery to be analysed both temporally and spatially at the level of individual stations. This dissertation addresses that gap by applying recovery rate measures and spatial clustering methods to the entry and exit counts of all stations, comparing patterns across fare zones and over time.

The overarching research question is:

How have commuting patterns on the London Underground changed in the aftermath of the COVID-19 pandemic, and to what extent can these changes be distinguished from pre-existing trends? Furthermore, do the recovery rates of stations exhibit spatial patterns, and how have these patterns evolved between 2021 and 2023?

From the perspective of forced experimentation, it is hypothesised that the pandemic would leave a long-lasting imprint on Underground ridership, with disruption observable between 2021 and 2023. In terms of geography, central London stations were expected to recover faster than outer areas, reflecting their higher density of population, workplaces, and social activity. With respect to travel purpose, commuting demand is expected to recover more slowly than all-purpose ridership, due to the persistence of remote and hybrid working arrangements. Together, these expectations framed the study as an investigation not only of how quickly ridership returned, but also of whether recovery followed systematic temporal and spatial patterns.

To examine these questions, this dissertation adopts a two-step approach. First, it defines recovery rates relative to a 2019 baseline, distinguishing between annualised ridership (all trips) and weekday commuting ridership (Monday–Thursday trips as a proxy for commuting). This enables temporal trajectories of recovery to be traced for each station. Second, it applies spatial clustering techniques, including Global Moran’s I, Local Moran’s I, and the Getis-Ord Gi* statistic, to identify whether stations with similar recovery profiles were spatially concentrated, and how these clusters changed between 2021 and 2023. By combining temporal and spatial perspectives, the analysis provides a fuller understanding of how disruption and recovery unfolded in the Underground network.

The remainder of this dissertation is structured as follows. Section 2 reviews literature on forced experimentation, transport disruption, and post-pandemic mobility, situating this study within broader debates and identifying gaps. Section 3 introduces the methodology, including the definition of recovery rates, use of fare zones, and application of clustering techniques. Section 4 presents results on pre-pandemic stability, temporal recovery, and spatial clustering patterns. Section 5 discusses these findings in relation to existing studies, reflecting on both consistencies and contradictions. Section 6 concludes by summarising contributions, highlighting limitations, outlining policy implications, and suggesting directions for future research.