Utility Week

Operations & Assets UTILITY WEEK | 24TH - 30TH NOVEMBER 2017 | 19 Market view A lthough they belong to an industry that offers an almost endless supply of data points, few UK utilities can truly say they are getting maximum value from the data they collect. This is not because utilities don't appre- ciate the potential impact data can have on their business. In a recent survey by Accen- ture, more than 90 per cent of utility execu- tives recognised it was important to buy or build advanced analytical and statistical tools and assure the availability of data sci- entists in the company. Yet despite this, a 20/20 report by PwC found only 23 per cent of UK utilities had placed most reliance on data for their last big decision. Why the disconnect? Too oen, utilities are still sold plug-and-play data applications that do nothing more than throw up a few fancy – and meaningless – numbers on a col- ourful dashboard. It is essential for analytics vendors to not only help utilities collate the analysis, but also help them to understand it and to verify its value in order to cement this growing belief in the power of data. The first step to finding value in data is to focus not on data processing, but rather on understanding data and what affects it. Too oen, data is treated as separate from the outside world. This is simply not the case. While data scientists calculate and provide predictions, interventions may be taking place that influence their accuracy. A simple way to demonstrate this is to use a false positive and true positive example of why data analysis cannot be treated in iso- lation. If a statistical model predicts that an asset will fail, and it does indeed fail a few months later, this is termed a "true positive". The prediction was correct. In the case that the model predicted a failure, but no failure took place, this is termed a "false positive". In the case of false positives, simply put, the prediction was wrong. Yet it's important to note the reason for this predictive failure could be an external intervention. For exam- ple, it's highly unlikely for a tree to fall on an overhead line if a vegetation trimming exercise has been performed preventing the tree from doing any damage. While the crew increased the grid's stability, it "damaged" the statistics by creating a false positive. As a full analogy, in the case of a "true positive", the model predicted that a certain asset would fail and, as expected, it does fail. The prediction was correct. But again, it's important to ask whether the reasons for failure match those provided by the statisti- cal model. In this situation a freak storm or other environmental anomaly could have damaged the asset and without this happen- ing the asset wouldn't have failed. In this instance the prediction would have failed and it was actually the freak storm that saved the statistical model's prediction. What these examples are intended to show is that finding value in data doesn't stop with predictions, and in order to find true value in data analytics, utilities need to take multiple variables into consideration, spend more effort, have a thorough look into the whole system and bring true domain know-how to the table. In short, utilities need to understand the causation that sits behind "true positive" and "false positive" data results to effect real change to the grid. These simple "dos and don'ts" can help support those undertaking this effort: • Don't believe that a single tool can help you solve your problem. Siloed applica- tions and platforms will just provide you with meaningless – and consequently expensive – numbers. • Do support your in-house experts by seeking the advice and guidance of external practitioners from analytics and data science domains to secure the most insightful results. • Don't trust one-dimensional outcomes. Your grid interacts and is affected by the real world. Balance your predictions against what your team is doing day-to- day for the best results. • Do take the time to understand your data. You'll only use it if you know what it says – and doing so could have a serious impact on your operational efficiency and business performance. Franz Winterauer, vice president, energy insight, EMEA, Omnetric Group Data: true or false? Utilities need to understand "true positive" and "false positive" data results to effect real change to the grid, says Franz Winterauer. T he Utility Week HR Forum I attended in September gave a good flavour of current concerns across the utilities sectors: skills gaps and how to fill them; aligning recruitment with national strategy; embedding female retention; the impact of technology and the need for change; and developing the new generation of engineers. These are all really important issues, and they exist as a thread throughout everything we do. However, dare I say it, everyday life in human resources can involve a good deal of fire-fighting as well as strategising. When you are beginning to feel you might be losing the plot, it is important to be abso- lutely sure you are doing the basics right. Focusing resource and effort on the key HR functions that touch and influence the lives of employees is essential. By concentrating on doing the basics right you can also drive forward the changes you want to improve key business indicators such as staff retention, a reducing skills gap, and operational productivity. So what are the basics? In my view they are recruitment, induction, review and appraisal, training, personal development and continuity plan- ning, and communication. There was discussion at the Forum about the need for honesty and openness during the recruitment process. This is especially the case in the wastewater sector. Being a drainage engineer is going to be a challenge if you're squeamish about bodily functions or smells. But you can emphasise the strong customer service element to the role, something we, as a sector, probably do not do enough. We will keep doing the basics. But that does not mean we will not be seeking to accommodate more of the broader range of issues we know are having a big impact on our business. Debi Bell, head of HR services, Lanes Group Pipe up Debi Bell "Focus resource and effort on the key HR functions that touch the lives of employees." "It is important to be absolutely sure you are doing the basics right"

• Cover