Internet of Things Means Everyone Wants a Data Scientist

 By
 Kachina Shaw

The potential value in the Internet of Things (IoT) is bringing to a fever pitch the focus on data as one of the enterprise’s most valuable assets. Clearly, those who carefully collect, transform, analyze, model and report on IoT data are seeing their influence rise. As much of this work is settling around the data scientist role, I talked with Don DeLoach, CEO of Infobright, provider of an analytics database platform, about what data scientists are being asked to do now, and how those responsibilities around IoT data might change in the near future.



DeLoach says it’s definitely early days when you look at what data scientists are being asked to examine:

“Look at the progress of the Internet of Things. Most, probably 95 percent, of the focus is on the closed loop message response systems that make up the use cases: service models for capital equipment, focus on specific silos, alerting to problems, not having to send service professionals out when they’re not needed, or information like temperatures in machines, or lighting levels that are appropriate for time or conditions. It’s grabbing a message off a sensor, and then determining whether an action is needed. We’re at an early stage.”

Where data scientists are wielding more influence, and adding more value, DeLoach says, is in incorporating data from both inside and outside the IoT environment:

“… gathering information on closed loop message response, beginning to realize the value in storing not only more data from silos BUT storage and cleansing of data across silos, including beyond the IoT environment. Now you’re using a massive amount of data and can leverage the utility of that data across many dimensions, to discover patterns. The work of the data scientist will be manifest in enhancing the value of silos, enhancing enterprise applications like CRM, capacity planning or the service analysis stack.”

One way to look at it is that the data scientist aims to answer three basic questions, says DeLoach:

1. What is going on? This is traditionally where business intelligence tools are focused.
2. Why is it going on? This gets into forensic triangulation, identifying patterns, or finding the proverbial needle in a haystack. At this point, the data scientist has put the data into a state to interrogate it.
3. What can we predict? Predictive models feed the execution engine, alerting systems and work flows, and are now a piece of the architecture. As they mature, they will transform into much more sophisticated, richer models, which means the organization can take better actions.

Two industries, says DeLoach, that provide illustrations of where we’re going with all this are the telecom industry, which takes data generated from operational support systems and machine-generated data, compresses it and moves it all to a central computing facility for analysis on a continual basis, and the financial trading industry, which is forever collecting vast amounts of data that goes into related factors, like weather and demographics. The collected data can be mined and interrogated, then baked into execution algorithms. These are now operating completely without human intervention, explains DeLoach, which is similar to what we’re moving toward in IoT.

I wanted to know where DeLoach thought the data scientist would best fit into the organizational structure, and his thinking is that before we can answer that question, we should consider that the IoT is actually forcing a change in the org chart.

“I see a rethinking of traditional structures. The accommodation of the IoT is forcing together the operational side and the tech side of the business. Most likely, the data scientist will be found in the middle, where classically we might see the business analyst.”

In addition, in research he recently conducted on salaries for data professionals, DeLoach says the fact that the skill sets for data scientists are not “confined to mathematics, but include a deep understanding of technology and architecture and how data gets brought together,” the average salary for statisticians is $90,000, while the average for data scientists falls between $135,000 to $140,000.

Finally, I asked about how best to make the case for this position, whether as a staff member or service provider/consultant. Begin, says DeLoach, with how developed upper management’s understanding of the IoT is:

 “It starts with trying to ensure that leadership understands the potential value of the IoT in the first place. Right now, the whole focus is on the front end, that is where most people’s heads are. But leadership needs to step back and appreciate how much value is in the data. They can ask themselves, what can I gain if I leverage this data? Increased market share, increased revenue, decreased operating expenses, a bottom line return to shareholders? I could talk all day about the higher-level benefits to mankind, but for now let’s keep the focus here. The IoT is an avenue to gain ground, which companies need to get sooner rather than later, as a competitive advantage. Later, it will be table stakes.”

Fresh off the Mobile World Congress in Barcelona which, DeLoach says, was “tantamount to 90,000 people getting an IoT conference,” he’s thinking of other changes that the move to the IoT and data science will bring.

“Supply and demand says that as the IoT evolves, data scientists become more prominent. At the same time, tool sets will be developed that will make everyone at least partly a data scientist. The data scientist won’t go away, but tool sets and the ability to perform some functions will be available to more of us. Think about in the early days of the automobile. When you bought a car, you also hired a chauffeur. You needed a qualified operator for that asset – the car. As the asset becomes mainstream, the market demands that element be simplified, and owners are able to drive their own cars. We’ll see these advances in programming and visualization tools that will allow this change.

Look, I have seven kids, and I am telling them to go into data science. Soon, everyone will want one.”