Help us use predictive analytics to improve safety in the workplace

We want to understand whether analytics and data science can be used to help prevent major health and safety incidents in our working environments. How can we use big data and data analytics to help us predict pitfalls and thereby target mitigation activities to reduce key health and safety risks?

We're looking for a team of data experts who can work with us on developing an analytic model which uses patterns and relationships in data to identify key causal factors and indicators of harm. If we know the risks we can take action to prevent incidents more effectively – bringing us closer to our Zero Harm ambition.


Challenge details

EDF Energy is the UK entity of EDF Group and consists of several Business Units with activities ranging from Power Generation (through Nuclear, Coal and Gas stations) to Metering and Customer facing Contact Centres.  This project will focus solely on the activities of EDF Energy’s two coal-fired power stations in Nottinghamshire - Cottam and West Burton A.

Incident reports at EDF Energy are categorised into 3 levels, “Incidents (I)”, “Serious Incidents (SI)” and “Very Serious Incidents (VSI)”.  This project is aimed at identifying the key casual factors of SIs.  Current analysis of such incidents mainly consists of manually making use of existing incident reports and near-miss data.

Spreadsheet-based reporting tools have been developed in-house to conduct basic slicing and dicing of data to identify trends and to display information through changing key variables. However, this is high-level with no linking of alternative data sets beyond incident reports.

Beyond this, more detailed analysis is carried out which is reliant on human business expertise rather than predictive analytics and software modelling to look for deep correlations and causal factors.

There is a wealth of information relating to health and safety performance data from the 2 power stations. Five core sources of data have been identified which are likely to be key to the modelling, these are:

  • Incident Reports
  • Dip Check Reports
  • Safety Observations
  • Audit Reports
  • Inspection Reports

These data sources contain a wealth of historical data, incident data alone dates back several years with thousands of records.

The data sources outlined above contain both structured and free text data and are not an exhaustive list of the information held by EDF Energy relating to health and safety. Other data may be available upon request to support the modelling approach such as card swipe reports or staff demographics.

Successful delivery of this ‘proof of concept’ exercise will see EDF Energy gain new insights into the factors which contribute to the occurrence of serious incidents.


guages in power station

Pitch to win a £20,000 contract

The winner from the pitch day will win up to £20,000 contract to develop a proof of concept for a predictive data analytics solution for EDF Energy. The contract awarded will require the winners to come and work for a few days at our coal-fired power stations in Nottinghamshire to get access to the databases.

If EDF Energy is happy with the work from the proof of concept contract, then a potentially much larger contract to further develop, implement and rollout a fully integrated solution may be awarded in 2017 (subject to standard procurement processes).

The winner will also be fast tracked into the last stages of the international EDF Pulse Awards in 2017 where the prize includes €100,000 (subject to a relevant category being available).

EDF Energy Pulse challenge key dates

Timeline of events - EDF Energy Pulse Awards



More challenges

Connected home and business

STEM - Getting children interested in science, technology, engineering and maths

Inspiring young people into STEM

Developing inspection and monitoring methods for nuclear reactors and offshore wind farms