Mining groups of factors influencing bus/minibus crash severities on poor pavement condition roads considering different lighting status

Abstract

Objective

This study employs a data mining approach to discover hidden groups of crash-risk factors leading to each bus/minibus crash severity level on pothole-ridden/poor roads categorized under different lighting conditions namely daylight, night with streetlights turned on, and night with streetlights turned off/no streetlights.

Methods

The bus/minibus data employed contained 2,832 crashes observed on poor roads between 2011 and 2015, with variables such as the weather, driver, vehicle, roadway, and temporal characteristics. The data was grouped into three based on lighting condition, and the association rule data mining approach was applied.

Results

Overall, most rules pointing to fatal crashes included the hit-pedestrian variable, and these crashes were more frequent on straight/flat roads at night. While median presence was highly associated with severe bus/minibus crashes on dark-and-unlighted roads, median absence was correlated with severe crashes on dark-but-lighted roads. On-street parking was identified as a leading contributor to property-damage-only crashes in daylight conditions.

Conclusions

The study proposed relevant countermeasures to provide practical guidance to safety engineers regarding the mitigation of bus/minibus crashes in Ghana.

Keywords:

• Developing country
• data mining
• severity
• crash
• bus
• lighting condition

Acknowledgments

The authors would like to thank the anonymous reviewers for their valuable comments, which helped improve the quality of the article.

Disclosure statement

None of the authors have any relevant competing interests to declare.

Funding

The authors report that there is no funding associated with the work featured in this article.