Recent technologies for recording and storing data, as well as advancements in data processing techniques, have opened up novel possibilities for urban planners to design a more optimal public transport network. This study aims to initially develop a robust framework for making an insightful understanding of already recorded and available data sets using machine learning approaches. This will give transportation planners a powerful framework to use great recorded datasets to understand the network better and make datasets more meaningful for transport planners. And then introduces an approach to use Machine Learning algorithms and extract hidden patterns for predicting financial loss during any crisis, which is a novel perspective and application. To do this, seven alternative machine learning algorithms were developed to predict ridership: Multiple Linear Regression, Decision Tree, Random Forest, Bayesian Ridge Regression, Neural Networks, Support Vector Regression, and k-Nearest Neighbors. The developed framework was applied to the available 10 years of historical recorded data from the blue bus line number 4 in Stockholm, Sweden. The best model, kNN, with an average R-squared of 0.65 in 10-fold cross-validation, was accepted as the best model. This model is then used to estimate the financial loss of the network during the pandemic in 2020 and 2021. Results reveal a decline of 49% in 2020 and 82% in 2021 in the studied line. Finally, the results were validated with a similar study that analyzed the ticket validations and passenger counts during the spring of 2020.
Machine learning framework to estimate ridership loss in public transport during external crises: case study of bus network in Stockholm
Machine learning framework to estimate ridership loss... M. Movaghar et al.
Eur. Transp. Res. Rev.
European Transport Research Review ; 17 , 1
2025-07-28
Article (Journal)
Electronic Resource
English
Regression , Public transport , Machine learning , Data-driven prediction , Ridership , Financial loss Statistics , Transportation , Automotive Engineering , Regional/Spatial Science , Mathematical Sciences , Civil Engineering , Information and Computing Sciences , Engineering , Artificial Intelligence and Image Processing
Using Behavioral Intentions to Estimate Public Transportation Ridership
British Library Conference Proceedings | 1991
Exploring the Factor of Public Transport Ridership
Springer Verlag | 2022
|WiFi Sensing System for Monitoring Public Transportation Ridership: A Case Study
Springer Verlag | 2020
|