Investigating the Pattern of Traffic Crashes Under Rainy Weather by Association Rules in Data Mining

Rainy weather and wet roads are considered hazardous conditions for driving. Countermeasures should be taken to reduce the risks of driving in such conditions, but measuring the added risk and key crash contributing factors under such conditions is very challenging. With a humid subtropical climate, the annual precipitation in Louisiana is about 64 inches, twice above the national average. Approximately 11% of total crashes in Louisiana happened during rainy weather, and nearly 25% of total fatal crashes happen in rainy weather annually. Reducing the number of crashes and crash severity is critical to the state “Zero Deaths Destination” highway safety strategies. The data mining technique is becoming immensely popular in dealing with huge dataset. It helps to identify the hidden patterns from a large and complex database, which is why these methods are being utilized in diversified areas. There are many data mining techniques that have been applied to traffic crash data analysis in the recent past. However, very little research work utilizes the association rules mining technique to discover knowledge from the traffic crash dataset. This data mining technique generates simple rules that introduce the association between different factors. This paper demonstrates how to apply this data mining methods to discover hidden patterns in rainy weather crash data with eight years of Louisiana data (2004-2011). No dependent variable is developed in this exploratory application contrary to many popular safety performance models. The findings of the research shows that ‘single vehicle run-off crashes’ is the most frequent item in rainy weather. This crash type is particularly associated with a few roadway features like ‘on grade-curve’ aligned roadways, curved roadways, and roadways with no streetlights at night. In rainy weather, Property Damage Only (PDO) and sideswipe (same direction) crashes are also significant in numbers. Moderate injuries are dominant in single vehicle crashes. Roadways with poor illumination are associated with straight level aligned roadways in rainy weather crashes. Young drivers (15-24) are vulnerable in run-off crashes when the roadways had poor illumination and are curve-aligned. The findings will help the highway authorities to determine countermeasure selection and safety improvement.

  • Order URL:
  • Supplemental Notes:
    • This paper was sponsored by TRB committee ANB20 Safety Data, Analysis and Evaluation.
  • Corporate Authors:

    Transportation Research Board

    500 Fifth Street, NW
    Washington, DC  USA  20001
  • Authors:
    • Das, Subasish
    • Sun, Xiaoduan
  • Conference:
    • Transportation Research Board 93rd Annual Meeting
  • Publication Date: 2014

Language

  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; References; Tables;
  • Pagination: 19p
  • Monograph Title: TRB 93rd Annual Meeting Compendium of Papers

Subject/Index Terms

Filing Info

  • Accession Number: 01519728
  • Record Type: Publication
  • Report/Paper Numbers: 14-1540
  • Files: TRIS, TRB, ATRI
  • Created Date: Jan 27 2014 2:34PM