The GPS of Data Analytics

We use detailed domain knowledge to help guide data analytics techniques to meaningful results. Without this knowledge, generic data analytics can yield results that are meaningless or misleading. This can include reporting correlations that:

  • Are already obvious and do not expand on the understanding of the system,
    • e.g. planes crash when they get too close to the ground;
  • Connect unrelated parameters where a cause-and-effect relationship cannot be made,
    • e.g. failure rates increase during full moons; and
  • Interpret missing data as evidence of low rates of occurrence,
    • e.g. pipes fail where no defects were recorded.

Domain knowledge can help avoid these pitfalls through:

  • Identifying what data is required, and what is not, to describe the system characteristics and performance;
  • Understanding the limitations of data sources, such as measurement accuracies, frequencies and locations, and how they impact what is recorded;
  • Applying engineering models to define how various parameters must be related, e.g. the nominal burst capacity of a pipe is a function of wall thickness, material strength and pipe diameter;
  • Constraining equipment specifications to the industry standards, design practices and regulations that were in use at the time of installation;
  • Accounting for maintenance, repair and replacement practices that change throughout the equipment operating life; and
  • Considering changes in operating environments that are caused by a maturing operating environment or by location changes for equipment installation.


Filter by
Post Page
Engineering Analysis
Sort by

Do you have a question or comment?

If you are interested in our engineering or testing services, want a demo of a software product, want to join a joint industry project, or are interested in a training course, we want to connect with you.

Please complete the contact form and we will respond within 48 hours.

Follow us on LinkedIn to keep up-to-date on the latest happenings at C-FER.