Driver Pattern Identification in Road Crashes in Spain

Sanjurjo-De-No, Almudena; Arenas-Ramirez, Blanca; Mira, Jose; Aparicio-Izquierdo, Francisco · 2020 · Crossref

DOI: 10.1109/access.2020.3028043

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates driver behavior patterns in road crashes in Spain, aiming to identify multivariate relationships between driver demographics (gender and age), collision types, injury severity, and specific traffic offences. The research is motivated by the need to move beyond univariate analyses, which often fail to capture complex behavioral interactions, to support more effective road safety policies and resource allocation by regulatory bodies like the Spanish Traffic General Directorate. The authors utilized a dataset from the Spanish Traffic General Directorate covering vehicle collisions between 2004 and 2013. After filtering for interurban passenger car collisions and debugging records, the final dataset comprised 145,904 drivers. The study employed Self-Organizing Maps (SOM), an unsupervised machine learning technique, to project 145,904 drivers from an 8-dimensional space of offence variables onto a 2D map, preserving topological relationships. This allowed for the clustering of drivers with similar offence profiles. To validate the SOM results, the authors compared them with standard K-Means clustering and used tests of proportions to analyze specific offence clusters. Missing data were handled by assigning a low probability value (0.25) to unknown offences, based on sensitivity analyses. The results revealed distinct multivariate driver behavior patterns dependent on gender and age. Male drivers were found to be more predisposed to committing offences, particularly when multiple offences occurred simultaneously, such as speed violations combined with alcohol or drug use. Female drivers were more prevalent in clusters with no offences or single, less severe infractions. The study also identified that injury severity and collision types are jointly determined by these demographic and behavioral factors. For instance, while men generally had higher crash rates, women showed higher vulnerability to severe injuries in certain age groups. The SOM analysis successfully identified clusters where specific combinations of offences, such as speeding and impairment, were significantly more common among male drivers, whereas physical defects were slightly more prevalent in males, potentially due to overestimation of driving skills. The significance of this research lies in its methodological contribution to road safety analysis by demonstrating the utility of SOM in uncovering complex, multivariate patterns that traditional statistical methods might miss. By identifying the relative importance and proportions of specific driver behavior patterns, the findings provide actionable insights for policymakers. These insights can facilitate the targeted development of prevention measures and the more efficient allocation of resources by traffic authorities, ultimately aiming to reduce crash rates and injury severity through better understanding of gender and age-specific risk factors.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• sex gender
• demographic disparities
• incidence prevalence
• comparative international
• crash typology
• vru crash typology

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Empirical Findings: crash risk outcomes, observational prevalence
• Methodological Resource: dataset resource