Predicting Delays for NJ Transit

LING CHEN

Planning & Urban Design

R Reimaging the City of Squares

Predicting Delays for NJ Transit Commuters

Oct 2023 - Dec 2023

Multiple predictive models for real-time train delay predictions up to 3 hours in advance, integrating spatial autocorrelation, serial, and time lag effects.

GitHub

Serving as the second largest commuter rail network in the United State, the NJ Transit spans New Jersey and the state to New York City. However, their delays are getting worse with more office workers returning, on the other hand, there are no interactive apps for commuters that can predict the immediate delays that may happen by chance or fixedly occur on their daily commute routes during the day. As such, they cannot foresee them instantly and mitigate accordingly.

In our project, we further investigate into the delay performance on NJ transit’s commuter rail routes and come up with some interactive & instant predictive strategies targeted at commuters within NJ Transit Commuter Rail Routes.

As such, Delay detective is designed with commuters in mind. The functions include the notifications of the estimated arrival time before scheduled arrival time, on-time & average-delay performance, as well as passenger feedback.

Type: Academic Work | Public Policy Analytics
Group work with Tianxiao Chen
Role: Background study and use case, Factor Analysis, Data Wrangling(Lag Effects), Model building
Location: New Jersey
Date: October, 2023 - December, 2023

Presentation:

Full Final RMarkdown:

Index Next →