Technology + Data

Data Science – How To Turn Data Into Actionable Business Decisions

As your organization increases its use of data, you must ask yourself, “Are we willing to change the way we make decisions?” This is one of the fundamental building blocks needed to gain added value. I refer to this as organizational change management, which must be led from the top and become part of the fabric of the organization.

To accomplish this, you must first develop a common language and understanding of what data science is. It is the ability to change data into actionable information that can improve customer satisfaction, operational efficiency and shareholder value. However, your definition of data science will vary based on your business strategies and expected outcomes.

As you begin to form a definition, remember that socializing data science will be key to your success. The utility industry is changing, and we must change with it or risk intermediaries taking our place.

We collect an immense volume of data from:
• generation,
• transmission and distribution sensors,
• advanced metering infrastructure (AMI),
• asset and outage management,
• financial reports,
• inventory and customer information systems,
• emails, and
• social media.

Now, what percentage of this data is being used to make decisions? Ten percent? Data is used in budgeting, customer service, scheduling, crew and supply-chain management, operations, outage management and asset management, to name a few. But what if we were able to double or triple the use of this data? Would we make different decisions and, perhaps, even monetize the results of our findings into something greater than we do today? What benefits could we provide to our customers through better customer service, operational efficiencies and outage restoration?

The more we know about something, the better positioned we are to add value. For example, if we knew which of our customers have electric vehicles, we could offer them programs that align with green energy. Knowing a customer used to mean knowing their energy usage and payment history. But that has changed. Using our customer data and combining it with external data, such as demographics, income, education and age, we now can yield new insights into our customers’ habits and preferences. Using data science could provide customers with new products and services that will allow them to make different choices on how they use their energy today.

In the area of generation, if we could better optimize the dispatch of generation resources along with market conditions, we could see increased efficiency and capture potential operations and maintenance savings. Likewise, what if we were able to view in real-time market prices, load, maintenance schedules, resource efficiency ratings, weather and water temperature — could we optimize our generation resources by even 1 percent? Depending on our generation portfolio, this could yield significant savings.

KEEPING UP WITH THE VALUE OF BIG DATA

As we move toward AMI, the Internet of Things and the Grid of Things, we anticipate gaining new insights on operational efficiency and restoration timeliness that will help control costs and increase customer satisfaction. Traditional, back-office analytics will be considered too slow to provide some of this value. Moving forward, we will see edge analytics used to capture low and high voltage, overloaded transformers, and even service interruption and power theft. The possibilities are endless. For example, consider having visual information of your operations center and the ability to dispatch the right crew with the right material to address adverse conditions on your distribution system.

Data is the currency of the 21st century, and those who invest in understanding its value will do better than those who fail to do so.

Likewise, we can use analytics from meters to improve customer restoration and to better understand load profiles, thereby, offering customers improved products and services. By providing customers with near real-time data, we can also encourage them to invest in home energy monitoring and automation to better understand their usage patterns. This includes hourly, daily and seasonal trends. Providing our customers with this information will allow them to make different choices about how they consume and pay for our services. We can also use external data to predict future weather and economic forecasts, which can help steer customers toward home improvements, such as furnace, air conditioning, appliance and lighting upgrades. We could also use these predictions to provide customers with service offerings, such as home energy audits, furnace cleaning and filter replacements, conversion to gas appliances, and solar generation and battery storage options. Combining customer information with external data, we can determine offerings around on-bill financing and low-income programs. Turning data into actionable information will increase customer satisfaction by allowing them to make choices that matter most.

How data is captured, validated and secured is one of our biggest challenges. If our business processes do not capture the right data, and if that data does not have good quality controls around it, we won’t know if we can rely on the information collected. This is known as data governance. Technology is an enabler for data governance, but more importantly, we have to ensure that our business processes support the collection and completeness of our data. Identifying the data that is most valuable to our interests is a key element of data governance. Since it is unrealistic to govern 100 percent of our data, we must identify the data we need that enables us to make better decisions, and focus on the systems that capture this data.

DEVELOP DATA SCIENCE

Once you have adopted a common language for data science, you can begin to define what, how and where you want to use your data. For example, begin by identifying all current electric customers in proximity of your gas distribution network, and then enrich this data set with external data. This includes information such as the year the customer’s home was built, whether the home is vented for forced air, whether the customer is a homeowner or a renter, and their income bracket. This information will allow us to provide customers with alternative energy choices for heating, cooking and hot water heaters.

Data is the currency of the 21st century, and those who invest in understanding its value will do better than those who fail to do so. Data science is a journey and we cannot expect instant results. Conversely, we must stay focused on improving the quality and availability of our data, while simultaneously looking for quick wins to keep the momentum going. Data science will change the way we do business, and we can either embrace this change or stand by and watch while others move ahead.

About the Author

Patrick Dever, Avista Utilities
Patrick Dever is the chief data strategist for Avista Utilities and is responsible for Avista’s enterprise data and information strategies. He has 16 years’ experience in software management of energy trading and risk management, utility operations, business intelligence and back-office systems.