Revealing hidden fall accident patterns in Indian construction using association rule mining.

Background: Falls from height are a leading cause of fatalities worldwide, yet the lack of systematic accident data has limited the development of targeted prevention strategies in India. This study aimed to identify patterns and causal factors associated with fatal fall accidents in the Indian construction industry between January 1, 2019, and March 31, 2025, using association rule mining. Methods: Owing to the unavailability of official government accident records, 126 fatal fall cases were collected from three major national newspapers. The cases were analyzed using the Apriori algorithm to identify frequent associations among accident-related variables. Strong association rules were evaluated across six consequence categories: causal factors, workers' state of origin, location of fall, occupation, state/region, and age. Results: The analysis generated 148 strong association rules. The absence of fall protection emerged as the dominant causal factor, consistently observed across most accident scenarios. Workers aged 18-25 years, general laborers, and migrant workers from Uttar Pradesh and Bihar were disproportionately represented among fatal fall victims. Scaffold-related and structure-related falls were the most common accident locations and were strongly associated with missing or improperly used fall protection. Maharashtra, Delhi, and Haryana were identified as the highest-risk regions. Conclusion: The findings show that even in data-scarce environments, systematic analysis of publicly available information can provide actionable safety insights. Fall prevention efforts in India should prioritize the provision and proper use of fall protection equipment, focus on young migrant workers and general laborers, strengthen enforcement in high-risk states, and establish a mandatory national accident reporting system to support evidence-based safety policy.