Data-driven customer segmentation: Assessing disparities in COVID impact on public transit user groups and recovery
COVID-19 triggered an unprecedented global lockdown and severely dampened public transit ridership, which was down 62% year-on-year across the U.S. through Q4 2020 [1]. Beyond these stark headline figures, more granular views of whose transit ridership patterns changed and how are needed to aid cash...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Published: |
Massachusetts Institute of Technology
2022
|
| Online Access: | https://hdl.handle.net/1721.1/138908 |
| _version_ | 1826216679886028800 |
|---|---|
| author | Luo, Rachel Li-Jiang |
| author2 | Zhao, Jinhua |
| author_facet | Zhao, Jinhua Luo, Rachel Li-Jiang |
| author_sort | Luo, Rachel Li-Jiang |
| collection | MIT |
| description | COVID-19 triggered an unprecedented global lockdown and severely dampened public transit ridership, which was down 62% year-on-year across the U.S. through Q4 2020 [1]. Beyond these stark headline figures, more granular views of whose transit ridership patterns changed and how are needed to aid cash-strapped transit agencies in understanding both the operational and equity impacts of COVID-19 and assessing possible recovery strategies [2].
This thesis examines these questions in the Metro Boston region by applying k-means clustering to smart card data from the Massachusetts Bay Transportation Authority (MBTA). We empirically determine customer segments based on passengerlevel pre-pandemic transit ridership patterns during January 13 - February 16, 2020, using data from 22.6 million trips by 1.5 million passengers. We then trace how COVID-19 produced differential churn rates and travel behaviour modifications among these distinct passenger groups. We find that COVID-19 induced churn among rail commuter segments key for supporting MBTA fare revenues, while bus riders and those who frequently rode rail off-peak—groups that covered the majority of reducedfare and vulnerable passengers—were most likely to continue using the system.
Our findings suggest that in the near term, the MBTA can support a ridership and revenue rebound by working closely with large employers involved in the MBTA "Perq" corporate pass program to plan for reopening. This can also position the MBTA to better gauge the need to redesign or reprice Perq to offer greater flexibility for workers who may be adopting remote work longer term and therefore commuting less frequently to the office. Further, our analysis reveals consistency in ridership patterns among bus passengers even during crisis times. In the medium term as the MBTA considers network redesigns to meet post-pandemic travel needs, existing plans for bus upgrades do not necessarily need heavy modification because COVID-19 did not completely redefine these passengers’ transit usage patterns. This gives a base level of certainty for the MBTA’s planning process, as it seeks to track and shape the uncertainty that COVID-19 has brought to demand on the rail side of its network. Finally, by supplementing our quantitative analysis with an overview of COVID-19 responses by other major U.S. transit agencies, we suggest that the MBTA can better weather future emergencies like COVID-19 by making longer-term efforts to shift its operating revenue mix away from volatile fare revenues towards more stable and resilient revenue sources such as sales and property taxes, and complementing this with sustainable financial management.
The framework offered in this thesis for dissecting passenger ridership behavior and tracking passenger churn and cluster-switching can be applied to other transit agencies to detail either background ridership behavioral changes in normal years or rapid step-changes during a mobility crisis. Understanding passengers’ ridership demand at the cluster level can inform both immediate actions that transit agencies can take to enable recovery, as well as support network redesign and long-term resilience. |
| first_indexed | 2024-09-23T16:51:42Z |
| format | Thesis |
| id | mit-1721.1/138908 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T16:51:42Z |
| publishDate | 2022 |
| publisher | Massachusetts Institute of Technology |
| record_format | dspace |
| spelling | mit-1721.1/1389082022-01-14T03:00:42Z Data-driven customer segmentation: Assessing disparities in COVID impact on public transit user groups and recovery Luo, Rachel Li-Jiang Zhao, Jinhua Aloisi, Jim Moody, Joanna Massachusetts Institute of Technology. Department of Urban Studies and Planning COVID-19 triggered an unprecedented global lockdown and severely dampened public transit ridership, which was down 62% year-on-year across the U.S. through Q4 2020 [1]. Beyond these stark headline figures, more granular views of whose transit ridership patterns changed and how are needed to aid cash-strapped transit agencies in understanding both the operational and equity impacts of COVID-19 and assessing possible recovery strategies [2]. This thesis examines these questions in the Metro Boston region by applying k-means clustering to smart card data from the Massachusetts Bay Transportation Authority (MBTA). We empirically determine customer segments based on passengerlevel pre-pandemic transit ridership patterns during January 13 - February 16, 2020, using data from 22.6 million trips by 1.5 million passengers. We then trace how COVID-19 produced differential churn rates and travel behaviour modifications among these distinct passenger groups. We find that COVID-19 induced churn among rail commuter segments key for supporting MBTA fare revenues, while bus riders and those who frequently rode rail off-peak—groups that covered the majority of reducedfare and vulnerable passengers—were most likely to continue using the system. Our findings suggest that in the near term, the MBTA can support a ridership and revenue rebound by working closely with large employers involved in the MBTA "Perq" corporate pass program to plan for reopening. This can also position the MBTA to better gauge the need to redesign or reprice Perq to offer greater flexibility for workers who may be adopting remote work longer term and therefore commuting less frequently to the office. Further, our analysis reveals consistency in ridership patterns among bus passengers even during crisis times. In the medium term as the MBTA considers network redesigns to meet post-pandemic travel needs, existing plans for bus upgrades do not necessarily need heavy modification because COVID-19 did not completely redefine these passengers’ transit usage patterns. This gives a base level of certainty for the MBTA’s planning process, as it seeks to track and shape the uncertainty that COVID-19 has brought to demand on the rail side of its network. Finally, by supplementing our quantitative analysis with an overview of COVID-19 responses by other major U.S. transit agencies, we suggest that the MBTA can better weather future emergencies like COVID-19 by making longer-term efforts to shift its operating revenue mix away from volatile fare revenues towards more stable and resilient revenue sources such as sales and property taxes, and complementing this with sustainable financial management. The framework offered in this thesis for dissecting passenger ridership behavior and tracking passenger churn and cluster-switching can be applied to other transit agencies to detail either background ridership behavioral changes in normal years or rapid step-changes during a mobility crisis. Understanding passengers’ ridership demand at the cluster level can inform both immediate actions that transit agencies can take to enable recovery, as well as support network redesign and long-term resilience. M.C.P. S.M. 2022-01-13T19:05:19Z 2022-01-13T19:05:19Z 2021-06 2021-07-27T20:27:26.196Z Thesis https://hdl.handle.net/1721.1/138908 In Copyright - Educational Use Permitted Copyright MIT http://rightsstatements.org/page/InC-EDU/1.0/ application/pdf Massachusetts Institute of Technology |
| spellingShingle | Luo, Rachel Li-Jiang Data-driven customer segmentation: Assessing disparities in COVID impact on public transit user groups and recovery |
| title | Data-driven customer segmentation: Assessing disparities in COVID impact on public transit user groups and recovery |
| title_full | Data-driven customer segmentation: Assessing disparities in COVID impact on public transit user groups and recovery |
| title_fullStr | Data-driven customer segmentation: Assessing disparities in COVID impact on public transit user groups and recovery |
| title_full_unstemmed | Data-driven customer segmentation: Assessing disparities in COVID impact on public transit user groups and recovery |
| title_short | Data-driven customer segmentation: Assessing disparities in COVID impact on public transit user groups and recovery |
| title_sort | data driven customer segmentation assessing disparities in covid impact on public transit user groups and recovery |
| url | https://hdl.handle.net/1721.1/138908 |
| work_keys_str_mv | AT luorachellijiang datadrivencustomersegmentationassessingdisparitiesincovidimpactonpublictransitusergroupsandrecovery |