Investigation on the wrong way driving crash patterns using multiple correspondence analysis.

Area of Science:

• Traffic Safety
• Transportation Engineering
• Accident Analysis

Background:

• Wrong way driving (WWD) crashes are a significant traffic safety concern, often resulting in fatalities or severe injuries.
• Previous research relied on traditional statistical methods, which can be biased due to the rarity of WWD events.
• Distribution-free methods like Multiple Correspondence Analysis (MCA) offer an alternative approach without requiring prior assumptions.

Purpose of the Study:

• To identify key contributing factors and their associations in wrong way driving (WWD) crashes.
• To apply Multiple Correspondence Analysis (MCA) to analyze WWD crash data from Louisiana (2010-2014).
• To provide data-driven safety recommendations for targeted countermeasures.

Main Methods:

• Utilized five years of wrong way driving (WWD) crash data from Louisiana (2010-2014).
• Employed Multiple Correspondence Analysis (MCA), a distribution-free statistical technique.
• Visualized variable category associations using a low-dimensional proximity map.

Main Results:

• MCA successfully identified sixteen significant clusters of contributing factors for WWD crashes.
• Key factors identified include specific locality types, dark roadways without lighting, roads lacking physical separation, higher speed limits, inadequate signage, and older drivers.
• The proximity map effectively illustrated relationships between these variable categories.

Conclusions:

• Multiple Correspondence Analysis is a robust method for analyzing rare events like wrong way driving (WWD) crashes.
• Understanding the clusters of contributing factors allows for the development of targeted safety interventions.
• The study provides actionable recommendations for authorities to mitigate WWD crash risks.