A background in data science versus a background in utilities? How can utilities hire the right people for their analytics initiatives?
Nearly every utility has a data analytics initiative today, or at least the beginnings of one. Like other data-driven industries, utilities are learning how to leverage data to cut costs and increase revenue. The utility especially has operational and business challenges driven by customer movement to distributed power generation, the need to optimize distributed energy resources, the influx of outside competition, and the requirement for resilience in the face of increased natural disasters.
However, most utilities still do not have in-house data scientists, and rely on outsourcing to build out their data analytics initiatives. Like every other industry that wants to get the most from its data, data scientists are in high demand at utilities.
Outsourcing has brought its own issues of fit and expense, but early on was the only game in town. Utilities simply had little real-world experience with the amount of data being produced by the modernized grid. Just from the perspective of smart meters alone, utilities lacked the infrastructure and skill required to handle interval consumption data, plus all the new meter values, such as voltage, volt-amperes, temperature, and meter events.
Still, even if a utility has a successful outsourcing relationship for their data management tasks, utilities that don’t develop internal skill and depth to generate and use models needed to run their operations are giving away a huge chunk of their core business.
If a utility company wants to extract the ROI of digitized assets, they must be able to move beyond a basic grid visibility to truly predictive capabilities. While grid data is useful at the description and diagnosis of field problems, new forms of business value come to fruition when data is used for prediction and prescription. These opportunities include predictive maintenance, more accurate energy utilization modeling, risk identification, and severe weather impact predictions.
Just who are these data scientists?
Now that buy-in among utility technicians and utility workers seems to have improved, many of the most florid articles about data analytics focus on design problems, stakeholder engagement, and organizational change. Yet, very little information exists for how to define the role of the utility data scientist or the positions that must be filled to make serious progress on analytics projects.
Despite the academy making some progress in educating students to work with the tools of data analytics, it is not altogether clear if hiring nascent computer scientists, mathematicians, or statisticians is the right strategy to fill key data scientist roles at a utility.
An anecdotal survey of data science graduate programs indicates that programs can be expected to include training in probability mathematics, programmatic data handling, and visualization. These are crucial technical skills, but should utilities really be looking to hire highly-skilled generalists at the expense of those with existing employees who could be trained to develop these same skills?
Hiring generalists and up-skilling existing employees both can be important strategies in building out a fully-fledged analytics program. However, in making key data scientist hiring decisions, a more complicated hiring or training process is called for. A lead data scientist must have skill, experience, passion, and business acumen that is directly relevant to utility industry.
Look for the outlier
Truthfully, the best data scientists need not be code jockeys nor ninja warriors (though they may be both), but must always be highly skilled at critical thinking and have extremely good intuition. Nearly anyone with the right degree of interest and instruction can learn to do a statistical regression, but not everyone has the contradictory talents for creating and understanding complex mathematical models with the ability to deploy those models to solve nuanced business problems.
This is not to say that there is not an important role for those with technical skills, mathematical chops, and the ability to create charts, graphs, and visualizations from mountains of data. What is easy to overlook, though, is the importance of creating meaning from that data. That often requires the ability to construct a narrative or compelling story that can only be told with utility-business acumen that helps predict outcomes and mitigate risks. A top-notch data scientist knows their way around a model and is willing and capable of making an objective analysis, but they also know what matters and doesn’t matter to solving business problems.
There is no better teacher than domain-explicit experience, especially when it comes to interpreting data in the high-risk and challenging environment faced by utilities. A solid data analyst knows how to solve big data problems, but a high-value data scientist can tell you how the models work, why they were designed that way, what they will tell you and not tell you. Because they have a deep understanding for the problems they are trying to solve, they won’t create distraction or generate new forms of risk for the business—hire the data scientist who has demonstrated skills in system engineering, but also who naturally comprehends utility risks, challenges, and what will bring success.
