查看文章

The Washington DC Metrorail system operates 1,126 railcars on six different routes over 117 miles of track to support over 500,000 passengers each day. The Washington Metropolitan Area Transit Authority relies on their On-Time Performance (OTP) metric to determine how well their system is running and identify delays in the system. This paper utilizes passenger tap-in/tap-out and train movement data to create a predictive model of OTP for current passengers in real time. These predictions can be used by WMATA to improve performance and communicate delays to passengers more effectively. Our approach goes beyond predicting OTP of current in-flight passengers and uses RNN predictions of the future network state to make OTP predictions for passengers who have not yet entered the network. These empirical applications can be powerful for agencies and planners to assess and improve transit service performance using big data analytics and real-time predictions.