A large construction company had access to data on historic incidents and the conditions and situations surrounding them, but were struggling to extract actionable insight from this information. They wanted to discover what were the key drivers of accidents, and use this information to take preemptive action on future projects to reduce the risk of injury to the workforce.
By processing and aggregating available data into a structured dataset, Mind Foundry was able to surface a number of causal relationships between incident severity, type, and environmental conditions.
This analysis helped to quantify the safety impact of undertaking work at different times of year, with different partners, as well as other factors. These relationships allowed the client to be proactive in monitoring at risk individuals and projects more thoroughly in order to maximise care, and minimise risk.
Using this data combined with historic health and safety inspection scores resulted in models that performed successfully in predicting the outcome of health and safety inspections before they had taken place.
Accurate predictions meant it was possible for the customer to monitor current projects, and send inspection teams to those most at risk of scoring poorly on an inspection, thus catching and remedying potential problems before they caused injury.
Time of Year
correlates strongly with key injury types
Natural Language Processing + Sentiment Analysis could unlock hidden insights in overlooked data