Literature Review and Data Analysis on the Impacts of Prolonged Events on Transit System Ridership
Liu, Yining; Osorio, Jesus; Ouyang, Yanfeng
The COVID-19 pandemic has drastically disrupted transit operations and induced significant transit ridership losses worldwide. Given its unprecedented duration, magnitude, and scale, the long-term effects are still unclear. Despite the differences, we can learn from previous disruptive events, such as terrorist attacks and epidemics, in the past 30 years and draw qualitative and quantitative insights about public reactions, ridership recovery periods, and transit agency responses during and after those events. This study sought to understand ridership variations during the current COVID-19 pandemic and inform transit agencies’ future decisions. This project’s research team therefore reviewed the impacts of selected historical events. They observed the following: (i) that most of the reviewed incidents (except for the 9/11 attacks) did not impose prolonged post-event effects on transit ridership for more than one year; (ii) that executive orders (e.g., school closures), transportation services (e.g., intensified airport safety screening and rail station closures), public fear, media reports, and reduced tourism were frequently mentioned as key factors that impacted transit ridership; and (iii) that measures, such as sanitizing vehicles and facilities, improving communications with the public, and promotions and advertisements, can potentially help restore transit ridership. The research team also developed a modeling framework that integrated a Bayesian structural time-series model, a dynamics model for daily transit ridership loss, a prediction module, and ordinary least squares regression to study COVID-19’s effects on the Chicago Transit Authority’s rail ridership. The researchers undertook a model of ridership on the CTA rail system as a potential first step to modeling COVID-19’s effects on transit ridership in northeastern Illinois. The researchers have not modeled ridership on any other transportation mode in northeastern Illinois at this time. The statistical analysis showed that remote learning/work policies and executive orders had answered for most of the ridership loss. Socioeconomic and land-use characteristics could effectively capture their effects. However, these characteristics could not explain people’s different reactions to reported deaths and media attention. Different population groups may have reacted differently to policy decisions, but their responses to reported deaths and media coverage seem random and independent of sociodemographic factors.
COVID-19; Transit Ridership; Historic Events; Bayesian Structural Time Series; Dynamics Model; Forecast
via Illinois Center for Transportation https://ift.tt/3kdqOLX
August 28, 2021 at 07:48AM