Page 145 - Intelligent Digital Oil And Gas Fields

108 Intelligent Digital Oil and Gas Fields

4.1.2 Data Analytics: Descriptive, Diagnostic, Predictive,
Prescriptive, and Cognitive

In today’s data-intensive world, four types of analytics are available to help
companies better harness the value of information (VOI) and knowledge
hidden in data. Gartner provides an excellent example of classification of dif-
ferent types of analytics from the perspective of added value and related
complexity of implementation and use, where Laney (2012) distinguishes
between four main concepts of analytics (Fig. 4.2): descriptive, diagnostic,
predictive, and prescriptive.
• Descriptive analytics is an approach to help us understand and answer the
question “what happened?” during a given past period and verify
whether or not a campaign or an action was successful. It is based on sim-
ple parameters such as number of trials or repetitions of a certain job.
>90% of companies today (oil and gas and other industries) use this type
of very basic analytics. However, even when using simpler, descriptive
analytics, the data must be explored, visualized, comprehended, and

How can we
make it happen?
Prescriptive
What will
happen? analytics
Why did it Predictive Optimization
happen? analytics
VALUE happened? Diagnostic Foresight
What
analytics
Descriptive Insight
analytics
Information Hindsight Cognitive analytics

Difficulty
Fig. 4.2 Four different types of data analytics from the perspective of added value vs.
related complexity of implementation. The red box was added to capture predictive
and prescriptive types of data analytics that are frequently associated with the
so-called cognitive analytics. (Modified from Laney, D., 2012. Information, Economics,
Big Data and the Art of the Possible With Analytics. Presentation by Gartner Inc., https://
www-01.ibm.com/events/wwe/grp/grp037.nsf/vLookupPDFs/Gartner_Doug-%20Analytics/
$file/Gartner_Doug-%20Analytics.pdf.)