Accident prevention through machine learning at a construction contractor

Research Project, 2022 – 2025

The research goal of this project is to analyse data on construction site accident causes and prevention, by using machine learning and an accident causation system-based model to investigate early risk-warning signs. This project forms the final portion of May Shayboun's PhD degree (supervised by Christian Koch and co-supervised by Dimosthenis Kifokeris), and is building on top of Shayboun's work that has already been undertaken for the finalization of her licentiate thesis (Shayboun 2022). It is planned to achieve the research goal by utilising the already collected and available dataset of accident reports in Shayboun (2022), and developing a machine learning-based model for extracting and analysing reported accident causes and prevention measures. The project also entails developing and testing a research-based prototype of a machine learning system for occupational accident prevention in a contracting company. The project’s main focus is on accident prevention and understanding occupational accident causes in the contractor's context.

The project will contribute to increasing the knowledge about construction occupational accidents and their prevention. Moreover, it will also provide experience and value to improving construction production and delivering a case study of a tested machine learning process. It is important to find solutions that surpass the underreporting and other limitations of machine learning application for the construction industry, through improved methods of analysis, the utilisation of existing accident causation theories, and extracting knowledge from accident reporting.