How “Data Confident” are you before embarking on analytics journey?
image credit: © Pop Nukoonrat | Dreamstime.com
- Dec 2, 2020 5:38 pm GMTNov 30, 2020 1:51 am GMT
- 221 views
This item is part of the Special Issue - 2020-12 - Data Analytics & Intelligence, click here for more
Utilities have come long way in past few years to start leveraging digital tools & technologies to better manage the safety & reliability of their networks. There are massive deployments of smart meters, line sensors, condition monitoring devices, intelligent batteries, smart inverters, and list goes on. These investments are made in good faith and compelling business case that put consumers at the heart of it, by promising affordable & reliable electricity. However, more often than not, the expected benefits are not realized or to very limited extent.
What are the reasons?
So, what are the reasons that doesn’t let utilities reap the anticipated benefits. Data-led insights, is key contributor in almost every cost-benefits analysis, driving to achieve the benefits from most of “smart investments”. Talk about smart meters, IoT devices, asset intelligence or anything—unless the “data is modeled smartly” and not just any data modeling, the benefits are unlikely to be achieved. There are more factors that can further increase benefits, but in this article, let’s focus on “Data Confidence” and “Confident Analytics” that can be trusted & leveraged by business.
How to ensure benefits are achieved?
Let’s take a quick example- Senior asset management executive in Huntsville utility, aims to make better decisions on asset replacements and hence save money on operational costs. He/she carved out “big data analytics” plan and went to market to buy advanced analytics platform/ tool. Before the executive knows it, utility already have spent couple of million $$ and still exploring what are lowest hanging use-cases? There was similar conversation with a director at LA based large utility in 2016, where she was planning to roll out 2nd phase of big data analytics program- just to define the use-cases & related “data confidence” (which could have been done in first place) & story still repeats itself other places.
So, what does it take to ensure the money invested in analytics system/ tools returns your business benefits?
- Well defined use-case
- Data confidence on selected use case(s)
- Understanding of the data science and modeling
- Understanding of the business & operating market
- AI Translators
- Engineering & domain understanding
- Collective & Human Intelligence
- Business Integration of the selected use-case(s)
Data Confidence Framework
Amongst all the success factors for analytics journey, data confidence is cornerstone for ensuring the “right start” that leads to planned business destination (achieving business outcomes).
Let’s continue with above example, where asset executive planned to implement analytics system (any product, tool, platform) to predict the “remaining useful life” of key asset class. Below snapshot is taken from typical data confidence framework, that indicates at very early start whether the selected use-case is viable to be implemented with existing data available or utility need to invest in any further sensing or monitoring to grab the most important data.
One can clearly notice that data confidence is only 23% in below framework and selected use-case is unlikely to give desired outcomes, until highly relevant data fields are not made available by “smart data modeling”. This goes back to machine learning/ AI algorithms, as the very nature of these models is to learn from history and having “rightly modeled data” on plate is a pre-requisite to start making correlations/ pattern findings.
Once you have successfully modelled the data, which is residing in different systems bearing different format/ type, your data confidence will be much higher to make those business-critical decisions. In below snapshot, utility have made an investment to bring relevant data together from different systems, spreadsheets, files, handwritten notes etc and improved their data confidence to have higher success rate with their investments in AI/ML systems.
What next after your use case implementation is successful?
Once you have achieved the successful outcome of shortlisted use case, and its accuracy & effectiveness is appreciated by business, what’d you do next? Here comes the hard part, how to make this use-case as part of BAU activity. Business Integration is critical to ensure the long-term benefits and helping the business (asset) managers to make use of advanced data intelligence in their daily operations will open up far more opportunities for AI & machine learning.
*Author wishes to acknowledge the efforts of Abhishek Vinuth, whose inputs were valuable to complete this post.