Predicting travel demand of a bike sharing system using graph convolutional neural networks

TL;DR

The study tackles hourly station-level bike-sharing demand forecasting and proposes a gate graph convolutional neural network (GGCNN) that operates on dynamic station graphs while integrating trajectory, weather, and locational access data. By learning a time-varying adjacency and employing gated spatial-temporal learning, the method achieves $R^2 = 0.82$, $MSE = 0.37$, and $RMSE = 0.61$ on Chicago's Divvy dataset, outperforming classical econometric and multiple ML/DL baselines. The work demonstrates the value of dynamic spatio-temporal graphs and gating mechanisms for accurate, scalable bike-sharing forecasting, informing rebalancing and resource allocation in urban mobility systems. These findings highlight practical implications for city planners and illustrate how integrating external data enhances short-term demand prediction in complex transit networks.

Abstract

Public transportation systems play a crucial role in daily commutes, business operations, and leisure activities, emphasizing the need for effective management to meet public demands. One approach to achieve this goal is by predicting demand at the station level. Bike-sharing systems, as a form of transit service, contribute to the reduction of air and noise pollution, as well as traffic congestion. This study focuses on predicting travel demand within a bike-sharing system. A novel hybrid deep learning model called the gate graph convolutional neural network is introduced. This model enables prediction of the travel demand at station level. By integrating trajectory data, weather data, access data, and leveraging gate graph convolution networks, the accuracy of travel demand forecasting is significantly improved. Chicago City bike-sharing system is chosen as the case study. In this investigation, the proposed model is compared to the base models used in previous literature to evaluate their performance, demonstrating that the main model exhibits better performance than the base models. By utilizing this framework, transportation planners can make informed decisions on resource allocation and rebalancing management.

Figures (11)

• Figure 1: Correlation of travel demand between stations
• Figure 2: Graph convolution layers MLP: Multi Linear Perceptron; X: Features nodes of the graph
• Figure 3: Gate Recurrent Unit Cell structure
• Figure 4: Gate graph convolution neural network structure
• Figure 5: The framework of the proposed demand bike share prediction