A multimodal data framework for motorcyclist injury severity on rural undivided roads

Barua, Swastika; Dutta, Anandi K; Das, Subasish · 2026 · Scientific Reports

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 addresses the critical safety issue of severe and fatal motorcycle injuries on rural undivided roads, which disproportionately contribute to traffic fatalities despite lower overall crash volumes compared to urban areas. Motivated by the high vulnerability of motorcyclists in these settings—characterized by higher speeds, limited infrastructure, and longer emergency response times—the research aims to identify specific crash mechanisms and contributing factors. The authors highlight a gap in existing literature, noting that prior studies often analyze structured data or narrative text in isolation and fail to adequately account for unobserved heterogeneity in rural crash scenarios. To investigate these patterns, the researchers analyzed a comprehensive dataset of 12,753 motorcycle crashes from rural undivided roads in Texas between 2017 and 2023. The study employs a multi-method framework integrating machine learning, econometric modeling, and natural language processing. First, Cluster Correspondence Analysis (CCA) was used to identify underlying patterns and group crashes based on roadway features, timing, and environmental factors. Variable importance was assessed using XGBoost to select key predictors. Second, cluster-based Random Parameter Logit (RPL) models, including Mixed Logit and Correlated Random Parameters Logit variants, were estimated to examine how risk factors influence injury severity while accounting for unobserved heterogeneity. Finally, Latent Dirichlet Allocation (LDA) topic modeling was applied to crash narratives to uncover thematic trends and contextual details not captured in structured data. The findings reveal that severe and fatal injuries are primarily driven by high-speed loss-of-control events, specifically run-off-road and overturn crashes on both straight and curved segments. Intersections represent a distinct severity mechanism where inadequate lighting and turning or yielding conflicts significantly increase injury risk. Additionally, nighttime crashes on rural segments, particularly those involving fixed objects or animals, emerged as a high-risk scenario, reflecting the compounded effects of limited visibility, high operating speeds, and reduced reaction times. The integration of narrative analysis with econometric results validated these patterns, showing that context-dependent heterogeneity is obscured in conventional pooled models. The significance of this research lies in its methodological advancement and practical implications for traffic safety policy. By combining structured data analysis with narrative topic modeling, the study provides a more robust understanding of crash causation. The results inform a suite of policy interventions grounded in the Safe System Approach, recommending context-sensitive speed management, rural infrastructure upgrades, helmet use promotion, and improved emergency and wildlife response. These targeted strategies aim to address the specific mechanisms driving severe injuries on rural undivided roads, offering a more effective alternative to generalized safety recommendations.

Key finding

Severe and fatal motorcycle injuries on rural undivided roads are primarily driven by high-speed loss-of-control events and nighttime crashes involving fixed objects or animals, rather than intersection conflicts alone.

Methodology

dataset

Sample size: 12753

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. Discovered via author_sweep_intake on 2026-05-28.

Stage	Outcome	Tool	Model	Prompt	Attempts	Completed
discover	success	author_sweep	—	—	2	2026-05-28
archive	success	canonical_url	—	—	1	2026-06-04
extract	success	cached	—	—	3	2026-06-10
clean	success	clean	—	—	1	2026-06-04
chunk	success	chunk	—	—	1	2026-06-04
embed	success	embed	Qwen/Qwen3-Embedding-8B	—	1	2026-06-04
enrich	success	—	—	—	1	2026-05-28
promote	success	—	—	—	1	2026-06-04
summarize	success	llm	qwen3.6-27b-prismaquant	summ-v5	2	2026-06-10
tag	success	vector_similarity	—	—	15	2026-06-11
verify	success	—	—	—	2	2026-06-10

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

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

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