By Stuart Gillen and Marla Rosner
If you haven’t heard the term “prescriptive analytics,” you will soon. While it’s a conversation that’s just getting started, its implications for businesses in general, and wind operations specifically, are massive.
What is it? According to Gartner, prescriptive analytics is “the final frontier of analytic capabilities.” The best way to understand it is to look at previous fields of analytics to gain perspective on how prescriptive analytics is built on previous analytics methodologies.
Descriptive analytics is considered the first stage of analytics. This refers to using technologies such as statistical analysis to comb through historical business data and conduct post-mortem analyses on past events. Doing so helps analysts to determine what happened in the past and why. At this point in time, the majority of businesses employ descriptive analysis as part of their normal operating procedures.
Next is predictive analytics, which looks ahead instead. Rather than try to establish why something happened in the past, predictive analytics seeks to use similar techniques of data analysis, along with newcomer machine learning, to anticipate the likelihood of future events. This is a technique businesses in the wind sector are only now adopting, but one that is already making waves throughout the industry. In a business where unexpected failures and unscheduled downtime are both costly and dangerous, the ability to predict an asset failure before it occurs is a powerful tool.
Prescriptive analytics takes this one step further. Using machine learning algorithms, prescriptive analytics uses historical and current data to not only predict what will happen, but to pinpoint why it will happen, recommend potential plans of action, analyze the ramifications of each possible choice, and even help set a chosen plan into motion.
Prescriptive analytics needs to be powered by robust machine learning software, which allows it to adapt to changing conditions, understand unstructured data with natural language processing, and scale to any size of operation.
If predictive analytics is a boon, prescriptive analytics is a game-changer. A prescriptive analytics algorithm can integrate with existing systems, such as a ticketing or maintenance system. Based on, for example, the data from past tickets, the algorithm can not only tell you that a driveshaft is going to crack, it can also tell you why it’s cracking, when it will crack, where to order the new part needed, and how to install the new part. You can easily have new parts waiting and ready to go two weeks in advance and schedule the replacement at a time convenient for your operations thanks to the information from prescriptive analytics.
The idea of prescriptive analytics may be new, but it stands ready to transform whole industries.
