Flooding related traffic crashes: findings from association rules
DOI: 10.1080/19439962.2020.1734130
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 the contributing factors of flooding-related traffic crashes in Louisiana, addressing a critical safety gap where traditional crash databases lack specific filters for flood events. Motivated by the high vulnerability of transport networks to heavy rainfall and the significant loss of life and property—evidenced by 449 crashes and 22 fatalities between 2010 and 2016—the research aims to identify key patterns to inform targeted countermeasures. The authors argue that traditional regression models often overlook subgroup effects and lack interpretability for practitioners, necessitating a data-driven approach that captures complex associations without prior assumptions. To achieve this, the researchers collected seven years (2010–2016) of police-reported crash data from Louisiana. Since the database did not explicitly tag flood-related incidents, the team employed natural language processing (NLP) and text mining algorithms to identify crashes using keywords such as "flood," "flooding," and "flooded," followed by manual verification to remove redundancies. This process yielded a dataset of 449 flooding-related crashes. The study integrated this data with precipitation records from the National Oceanic and Atmospheric Administration (NOAA) and roadway attributes. The core analytical method was association rule mining using the Apriori algorithm, implemented via the R package 'arules'. This technique identified latent patterns by calculating support, confidence, and lift values for various itemsets, allowing the researchers to determine strong associations between crash characteristics, environmental conditions, and driver behaviors. The analysis revealed that rear-end collisions and driver violations are the most frequent attributes in the generated rules. Specifically, the strongest association rule indicated that crashes involving driver violations on two-way roadways with physical separation, occurring away from intersections, were highly likely to be rear-end collisions. Other prevalent factors included normal driver conditions, summer seasons, and high average annual precipitation (particularly 61–63 inches). The results showed that two-way roadways with separation often experience stalled water, contributing to these crash types. Additionally, crashes were predominantly property damage only (PDO), though injury crashes were also linked to violations and normal driver conditions. The study found that younger drivers (15–24 years) and specific posted speed limits (25–50 mph) were also frequently associated with these events. The significance of this research lies in its provision of a comprehensive, data-driven profile of flooding-related crashes, moving beyond general weather impacts to specific actionable insights. By identifying that driver violations and specific roadway geometries (two-way with separation) are strongly associated with rear-end crashes during floods, the study supports the development of targeted countermeasures. These may include improved warning signs, better water depth gauges, and public awareness campaigns to mitigate risky driving behaviors in flooded areas. The use of association rule mining demonstrates a viable method for extracting interpretable safety insights from complex, unstructured crash data, offering a practical tool for transportation safety planners to reduce fatalities and property damage in flood-prone regions.
Key finding
Association rule mining of Louisiana crash data identified that rear-end collisions, driver violations, two-way roadways with separation, and high average precipitation are the most frequently associated factors in flooding-related traffic crashes.
Methodology
dataset
Sample size: 449
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 | — | — | 7 | 2026-06-06 |
| 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.