Predicting risk

In the predicting risk work  the feasibility of deploying off the shelf predictive analytic tools for use in health and safety contexts has been explored. Taking the form of a pilot study with BAM Nuttall, a construction company, and two University of Oxford spin-off companies with expertise in semantic technologies and machine learning, namely Oxford Semantic Technologies and Mind Foundry.

 

    Aims and objectives

    The pilot study was designed to test a number of hypotheses:

    1. that it was possible to bring together a suitable training dataset built around routine health and safety data generated by organisations for use in a predictive analytic exercise
    2. that it was possible to build a predictive model able to predict health and safety endpoints of interest with a sufficient level of accuracy
    3. that it was possible to communicate the outputs of the modelling exercise in such a way that they added tangible value to health and safety targeting and actioning within an organisation
    4. that organisations without significant data science and data analytic expertise could deploy the tools and techniques developed independently and effectively, moving forward from the pilot

    Key benefits

    Demonstration that predictive analytic techniques can be successfully used in health and safety contexts to predict future health and safety endpoints as part of a strategy to promote more proactive, preventative health and safety practice

    Equipping organisations with the technical capabilities and providing them access to suitable analytic tools enabling them to exploit their own data resources more effectively, ultimately for the betterment of their health and safety performance