Unraveling urban multi-modal travel patterns and anomalies: a data-driven approach
Year: 2025
Type: article
Abstract: Urban transportation networks exhibit complex multi-modal demand patterns that vary across the city’s spatial terrain, making a comprehensive understanding of such patterns challenging. This research investigates the spatiotemporal dynamics of urban travel demand patterns using multi-source data from Lyon and Villeurbanne in France, integrating vehicle loop counts, public transport ticketing data, and bike-sharing usage. This work bridges data-driven travel demand analyses with spatial analysis, informing effective urban planning strategies. With a multi-modal approach, our study aims to uncover daily demand patterns, detect and classify anomalies, and explore their connections to the socioeconomic and spatial characteristics of urban zones. Latent Dirichlet Allocation (LDA) is employed to identify distinct multi-modal daily travel demand patterns, while clustering is used to group anomalies based on their multi-modal dynamics. Regression analyses further inspect relationships between zonal characteristics and normal/abnormal travel patterns. Findings reveal complex links between travel patterns and zonal characteristics but hint at the significance of metro/tram accessibility in shaping a zone’s multi-modal demand patterns. A meaningful connection between the incidence of multi-modal anomalies and zonal characteristics was not found. The study introduces novel methodologies for extracting multi-modal and spatiotemporal travel patterns, as well as anomaly detection and classification.
Language: English
Subfield: Transportation
Field: Social Sciences
Domain: Social Sciences
Open Access status: gold
Estimated APC paid: $1,165
Funder: Agence Nationale de la Recherche
OpenAlex ID: W4408896506