Finding the Right Personnel Resources to Analyze Big Data
IT and OT executives in the process manufacturing industries know there is tremendous value just waiting to be unearthed from their big data, but they're not always sure about what type of talent is required to turn this raw data into actionable information. Executives in other industries and commercial sectors with complex operations have similar challenges. In both cases, the three main choices are:
1. Finding and hiring data scientists. These rare, scarce and expensive individuals combine a high degree of IT expertise with deep knowledge of statistics and high-level math.
2. Instilling IT personnel with a high degree of OT expertise
3. Providing OT personnel with powerful software tools that are simple to use and don't require extensive IT support
This article will examine these three alternatives and recommend the best course of action.
The process manufacturing industries produce chemical, oil & gas, pharmaceutical, food and beverage, and other products. Power and water/wastewater facilities are also included under the process manufacturing industry umbrella, as are mining and other resource extraction activities. The process industry will be used as a basis for this discussion, but much of these observations will apply to other industries as well, and to some complex commercial operations such as logistics and telecoms.
Data scientists, often referred to as data citizens, are purported to exist in large numbers, but in reality can be very difficult to find. Once found, they tend to come with very high salary demands. Their unique combination of IT and math skills makes them very valuable, but in our experience, we have found no instances where this type of person has the required knowledge of process operations, and we suspect the same will be true for other complex operations.
Data scientists may prove useful quickest in sectors where detailed physical, chemical and/or biological training and experience are not required
Data scientists may prove useful quickest in sectors where detailed physical, chemical and/or biological training and experience are not required. These uniquely skilled individuals are often paired up with operational experts to solve complex problems.
The second option is to take existing IT personnel and train them in two areas: data analytics and operations. The advantage of this approach is existing personnel can be used, and their IT expertise usually allows them to pick up on data analytics concepts quite quickly. They have extensive experience with databases, and in general know how to work with software applications to analyze data and optimize operations, such as with computer network traffic.
But as with data scientists, our experience shows few if any IT personnel have the requisite knowledge of process operations. Therefore, extensive training would be required to bring an IT expert up to speed on process operations, and certain complex processes could remain outside their realm of understanding.
This brings us to the third alternative, providing OT personnel with powerful software tools that are simple to use and don't require extensive IT support. This is the approach we’ve seen work well with not only our internal staff but also with our customers’ staff.
In the process industries, companies like ours have extensive OT expertise, primarily in the form of engineers and scientists with many years of process operations experience. Our customers’ staff is similar, with the main difference being their higher degree of specialization in their particular plants and processes. In both cases, the requisite OT knowledge is in place in the form of experienced staff but is usually not accompanied by a high level of IT expertise within the same person.
The solution we’ve found most effective is to simplify the data analytics step—the transformation of raw data into actionable information—through the use of software apps designed to analyze data with solutions that are simple to use. For example, a pump app looks at data from sensors mounted on or near a pump and guides users through the required data analytics.
Because each app is simple to work with, it requires minimal training, even for those with very little IT expertise. This is done by breaking problems into many smaller pieces where root causes can be more quickly identified, along with clear actions for recovery. All of these pieces can then be rolled up into a simple dashboard view, providing a broader understanding of how the various parts connect together in a complex system.
For complex operations such as process plants and facilities, the likelihood of finding data scientists with the required level of OT skills to perform data analytics is slim. Training internal IT personnel to become OT experts presents its own challenges and is often not practical. The best alternative is to instead provide internal OT experts with simple yet powerful data analytics tools.