Predictive Safety Analytics Model for Early Detection of High-Risk Construction Activities

Abstract; The construction industry continues to experience a high incidence of occupational accidents due to complex work environments, dynamic project conditions, and the involvement of multiple stakeholders and high-risk activities such as working at heights, heavy equipment operation, and confined space tasks. Traditional safety management approaches often rely on reactive methods that investigate incidents after they occur, which limits their effectiveness in preventing future accidents. This review paper examines the development and application of predictive safety analytics models designed to enable early detection of high-risk construction activities before accidents occur. The study synthesizes existing research on data-driven safety management frameworks that integrate historical accident records, real-time site monitoring data, wearable sensor information, and environmental indicators to identify patterns associated with unsafe conditions and behaviors. Emphasis is placed on machine learning techniques, statistical risk modeling, and predictive algorithms that analyze large safety datasets to forecast potential hazards and support proactive decision-making. The review also evaluates the role of emerging technologies including Internet of Things (IoT) devices, computer vision systems, and digital twin platforms in enhancing predictive safety analytics within construction environments. By examining current methodologies, implementation challenges, and performance evaluation strategies, the paper highlights the potential of predictive analytics to transform construction safety management from reactive compliance-based systems into proactive risk prevention frameworks. The findings provide insights for researchers, construction managers, and safety professionals seeking to improve workplace safety outcomes through advanced data-driven risk prediction models.