Abstract:Big data provide massive samples and resources for exploring the operating rules of public transportation. This article proposes a method that combines multiple data sources to identify potential bus passenger flows, aiming to address the issue of insufficient identification accuracy with a single data source. First, the spatially weighted K‐means algorithm and improved DBSCAN algorithm are designed to partition traffic zones and residents' travel flow OD is extracted based on mobile phone signaling data. Second, using bus IC card data and vehicle trajectory data, a method for identifying bus passenger boarding and alighting stops based on spatiotemporal clustering is proposed and the bus passenger flow OD for each traffic zone is calculated. By comparing the resident travel flow OD with the bus passenger flow OD, we set a threshold for the potential bus passenger demand proportion. Finally, the analysis is conducted using actual data from a city in central China. The city is divided into 43 traffic zones, with the maximum bus passenger flow proportion between zones being 14.9%, the minimum being 5.0%, and the average being 7.2%. The initial threshold for the potential bus passenger demand proportion is thus set to 7.2%, and a sensitivity analysis is conducted by gradually decreasing the threshold in increments of 0.5% to 6.7%, 6.2%, 5.7%, and 5.2%. The corresponding potential bus passenger demand OD pairs between traffic zones are identified as 419, 358, 245, 151, and 51. Urban managers should focus on the 51 pairs with relatively large potential flows to gradually optimize and balance the development of the bus network based on actual conditions. The method proposed provides important theoretical and practical support for effectively optimizing urban bus networks. However, there are limited indicators for identifying potential passenger flows; in the future, more multidimensional indicators will be taken into consideration.

A Bus Passenger Flow Prediction Model Fused with Point-of-Interest Data Based on Extreme Gradient Boosting

Passenger Flow Prediction Using Smart Card Data from Connected Bus System Based on Interpretable XGBoost

A Vision-based Method to Estimate Passenger Flow in Bus

Study on the Prediction Model of Short-term Bus Passenger Flow Based on Big Data

A Real-Time Passenger Flow Estimation and Prediction Method for Urban Bus Transit Systems

Bus Passenger Volume Forecasting Model Based on XGBoost Integrated Learning Algorithm

Bus-Passenger-Flow Prediction Model Based on WPD, Attention Mechanism, and Bi-LSTM

Forecast and Early Warning of Regional Bus Passenger Flow Based on Machine Learning

A separate modeling approach for short-term bus passenger flow prediction based on behavioral patterns: A hybrid decision tree method

A hybrid ensemble forecasting model of passenger flow based on improved variational mode decomposition and boosting

An Efficient Approach for Identifying Potential Bus Passenger Demand Based on Multisource Data

Traffic Flow Prediction with Improved SOPIO-SVR Algorithm

Passenger flow prediction and management method of urban public transport based on SDAE model and improved Bi-LSTM neural network

A method for short-term passenger flow prediction in urban rail transit based on deep learning

Predicting Passengers In Public Transportation Using Smart Card Data

Passenger Flow Prediction in Bus Transportation System using ARIMA Models with Big Data

Attention Mechanism‐based Model for Short‐term Bus Traffic Passenger Volume Prediction

Multitype Origin‐Destination (OD) Passenger Flow Prediction for Urban Rail Transit: A Deep Learning Clustering First Predicting Second Integrated Framework

Exploring Bus Stop Mobility Pattern: A Multi-Pattern Deep Learning Prediction Framework

Exploring Human Mobility for Multi-Pattern Passenger Prediction: A Graph Learning Framework