Don't miss the new oil of the 21st century…

The 2.5 quintillion bytes of data produced worldwide each day

We must exploit this data
to increase our performance

Data Scientists can address this challenge:

    At Total, our Data Scientists are already exploiting these deposits

    Thanks to them, we are implementing innovative solutions across the Group:

    • Predictive maintenance

    • Risk management

    • Virtual assistance

    • Fraud prevention

    Perform predictive maintenance on our facilities

    SITE: Dalia FPSO & Girassol FPSO, oil facilities, Angola

    PROBLEM: high costs associated with predictive maintenance activities on our equipment (electrical systems and rotating machinery such as motors, pumps, turbines, etc.)

    SOLUTION: carry out predictive maintenance based on data analysis:
    - data collection via sensors attached to the equipment (approx. 150,000 values collected over a three-year period)
    - correlation between different data sources (physical measurements produced by sensors, such as vibrations, temperatures, pressure, combustion data, mechanical variables, etc.)

    - fault detection and predictive modeling of failure risks
    - real-time equipment monitoring on demand

    Anticipate leak risks due to equipment corrosion

    SITE: Leuna refinery, Germany

    PROBLEM: deterioration of equipment on our industrial sites due to corrosion

    SOLUTION: analyze and assess the impact of the following variables:
    - weather
    - product quality
    - equipment maintenance
    - human factor (decision-making and actions, training, etc.)

    RESULT: improved understanding of equipment corrosion sources (multiple causes) and a decision support tool for inspection

    Support our traders with intelligent virtual assistants

    PRODUCT: a virtual assistant developed by Trading & Shipping's Digital team

    PROBLEM: a situation of information overload that hinders quick decision-making

    SOLUTION: create a virtual assistant that understands natural written and spoken language and can execute orders using algorithms for:
    - machine learning
    - artificial intelligence
    - classification

    RESULT: a demonstrator currently under development that can:
    - show graphs (e.g., show me the Brent curve)
    - issue alerts about the evolution of market price trends (e.g., when Brent crosses a threshold)
    - predict shipping routes (e.g., track cargo and ship destinations)
    - and soon, summarize news articles to produce press reviews!

    Prevent fraudulent use of our GR Cards

    PRODUCT: the GR Card, issued to our B2B customers for access to services at our stations (carwash, maintenance, shopping, lubricants) and at toll gates, parking lots and for fuel removal

    PROBLEM: card duplication and production of false cards

    SOLUTION: process the records of toll gate transactions to develop:
    - a tool for the detection of abnormal behavior
    - a prediction model to anticipate counterfeited card numbers

    - model roll-out in France
    - 20% cut in fraud due to false card use

    Data Science is an exceptional opportunity for all of us

    Tempted by this new exploration?

    Discover our Data Science Challenge

    This challenge is open to all our co-workers who have a passion for data.
    From October 17, 2016 to February 1, 2017, connect to our platform to address the challenges of three Total professions.

    With the grand finale
    on February 1, 2017

    Hosted at our new innovation space: The Lab
    More information on the Challenge website

    So we can keep energizing life - and data - every day!

    Data Science Challenge - Total

    Exploit the new oil of the 21st century


    Photos: © Dalton Scott - CAPA Pictures - Total;
    © Schaff Philipe - Total; © Zylberman Laurent
    - Graphix Image - Total; © Perrin Guillaume - Total;
    © Blaise Bernard - Total
    Production: Storymakers; Total, © All rights reserved