Non-Stationary Time Series Model for Station Based Subway Ridership During Covid-19 Pandemic (Case Study: New York City)

TL;DR

This paper introduces a non-stationary, change point detection-based ARIMA modeling approach to analyze NYC subway ridership during COVID-19, capturing shifts in patterns caused by the pandemic.

Contribution

It proposes a novel piece-wise stationary time series model with data-driven change point detection to better understand ridership dynamics during external shocks.

Findings

Effective modeling of ridership changes during COVID-19 pandemic.

Identification of key change points in ridership patterns.

Enhanced understanding of temporal correlations in non-stationary data.

Abstract

The COVID-19 pandemic in 2020 has caused sudden shocks in transportation systems, specifically the subway ridership patterns in New York City. Understanding the temporal pattern of subway ridership through statistical models is crucial during such shocks. However, many existing statistical frameworks may not be a good fit to analyze the ridership data sets during the pandemic since some of the modeling assumption might be violated during this time. In this paper, utilizing change point detection procedures, we propose a piece-wise stationary time series model to capture the nonstationary structure of subway ridership. Specifically, the model consists of several independent station based autoregressive integrated moving average (ARIMA) models concatenated together at certain time points. Further, data-driven algorithms are utilized to detect the changes of ridership patterns as well as to estimate the model parameters before and during the COVID-19 pandemic. The data sets of focus are daily ridership of subway stations in New York City for randomly selected stations. Fitting the proposed model to these data sets enhances our understanding of ridership changes during external shocks, both in terms of mean (average) changes as well as the temporal correlations.