Predicting subway passenger flow and demand using big data and machine learning methods.

Authors: Xining Yang*, Eastern Michigan University, Yu Feng, University of Michigan, Hua Cai, Purdue University
Topics: Spatial Analysis & Modeling, Transportation Geography, Geographic Information Science and Systems
Keywords: human dynamic, big data, machine learning
Session Type: Virtual Paper
Day: 4/10/2021
Start / End Time: 9:35 AM / 10:50 AM
Room: Virtual 9
Presentation File: No File Uploaded

Understanding the quality and usage of public metro resources is important for schedule
optimization in the temporal dimension and route planning in the spatial dimension. Reliable subway
prediction is important for passengers, transit operators, and public agencies. This study
proposes novel machine learning methods to forecast the subway passenger flow and
demand combined with the big transit fair transaction data. Geovisualization is conducted to
effectively visualize the distribution of passenger flow over space and time. Empirical studies in
Wuhan demonstrate that the propose prediction model can effectively forecast the human
dynamic flow on utilizing the subway system under different external conditions.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login