Preliminary Analysis of Construction Work Zone on Roadways in Florida by Crash Severity

TL;DR

This study analyzes construction work zone crash data in Florida to identify key factors influencing fatalities and injuries, aiming to improve safety strategies through statistical modeling and machine learning insights.

Contribution

It provides a preliminary multilogit analysis of crash attributes in Florida work zones, highlighting significant factors and proposing machine learning applications for safety improvements.

Findings

Significant factors include crash type, weather, and work zone conditions.

High-risk counties identified for targeted safety interventions.

Machine learning can enhance driver alerts and safety management.

Abstract

Construction zones are inherently hazardous, posing significant risks to construction workers and motorists. Despite existing safety measures, construction zones continue to witness fatalities and serious injuries, imposing economic burdens. Addressing these issues requires understanding root causes and implementing preventive strategies centered around the 4Es (Engineering, Education, Enforcement, Emergency Response) and 4Is (Information Intelligence, Innovation, Insight into communities, Investment, and Policies). Proper safety management, integrating these strategic initiatives, aims to reduce and potentially eliminate fatalities and serious injuries in work zones. In Florida, road construction work zone fatalities and serious injuries remain a critical concern, especially in urban counties. Despite a 12 billion dollars infrastructure investment in 2022, Florida ranks eighth nationally for fatal work zone crashes involving commercial motor vehicles (CMVs). Analysis from 2019 to 2023 shows an average of 71 fatalities and 309 serious injuries annually in Florida work zones, reflecting a persistent safety challenge. High-risk counties include Orange, Broward, Duval, Hillsborough, Pasco, Miami-Dade, Seminole, Manatee, Palm Beach, and Lake. This study presents a preliminary analysis of work zone crashes in Broward, Duval, Hillsborough, and Orange counties. A multilogit model assessed attributes contributing to fatalities and serious injuries, such as crash type, weather and light conditions, work zone type, type of shoulder, presence of workers, and law enforcement. Results indicate significant contributing factors, highlighting opportunities to use machine learning for alerting drivers and construction managers, ultimately enhancing safety protocols and reducing fatalities.