Presented By O'Reilly and Cloudera
Make Data Work
22–23 May 2017: Training
23–25 May 2017: Tutorials & Conference
London, UK

Driving business value: Predicting piston ring failures in massive vessels

Mads Ingwar (Think Big), Eliano Marques (Think Big)
16:3517:15 Thursday, 25 May 2017
Secondary topics:  Deep learning, IoT
Level: Intermediate
Average rating: ****.
(4.50, 2 ratings)

Who is this presentation for?

  • Data scientists, analytics leads, and operations and maintenance leads

Prerequisite knowledge

  • A general understanding of maintenance and analytics

What you'll learn

  • Understand how data science was leveraged to plan ship engine maintenance by warning about potential piston ring failure


For the biggest shipping companies in the world, a daily challenge is to understand the ocean of sensor and engine data generated by vessels in order to help reduce and predict equipment failures. Being able to predict equipment failures ahead of time such that timely preventive actions can be taken to reduce unplanned downtime has significant potential. Eliano Marques and Mads Ingwar share a case study on how to leverage data science to plan ship engine maintenance by warning about potential piston ring failure.

Eliano and Mads walk you through how they used machine learning to predict engine failures and how they used raw sensor data to create actionable insights, covering the strategies they adopted and the feature engineering they developed in partnership with SME to combine sensor data with a multitude of unstructured datasets, as well as the different techniques they used to train and cross-validate the ML algorithms. Eliano and Mads conclude by outlining the business validation steps that they conducted to validate the use case and the refinements they made during deployment, including deep learning, to improve the process.

Mads Ingwar

Think Big

Mads Ingwar is the client services director at Think Big, a Teradata company, where he is responsible for leading consulting teams delivering data science and big data analytics combining Hadoop and Spark, the public cloud, and traditional data warehousing. Mads has a proven track record in using data science and big data to bring measurable added value to everything from startups to Fortune 500 companies. Mads has a strong background in pervasive computing and sensor-based technologies and holds a PhD from the Technical University of Denmark, where his research focused on machine learning and data analysis.

Eliano Marques

Think Big

Eliano Marques is the head of data science at Teradata International. Eliano has successfully led teams and projects to develop and implement analytics platforms, predictive models, and analytics operating models and has supported many businesses making better decisions through the use of data. Recently, Eliano has been focused in developing analytics solutions for customers around deep learning, AI, predictive asset maintenance, customer path analytics, and customer experience analytics across different industries. Eliano holds a degree in economics, an MSc in applied econometrics and forecasting, and several certifications in machine learning and data mining.

Comments on this page are now closed.


19/06/2017 15:27 BST

Hi Mads, can you share your presentation? I couldn’t find it in the speaker slides made available by the Strata Data Conference team.
Thank you