Sep 30, 2020
Community Spotlight: Innowatts provides AI-driven analytics for the power industry
Revolutionizing an industry with an AI-driven SaaS platform built on Apache Druid
Innowatts is a rapidly growing, AI-enabled SaaS platform that delivers near-real-time energy analytics and actionable business intelligence to utilities and retailers from around the world. Energy providers rely on Innowatts and the 40 million meters they are managing for the insights needed to be more predictive, proactive and connected, helping them unlock grid edge opportunities, increase customer value, and accelerate the transition to sustainable energy solutions.
We recently spoke with Daniel Hernandez, Apache Druid Engineer, about how Innowatts uses Druid to aggregate meter level data to create reports and recommendations for customers.
“Our customers gain insights about their grid and at the same time the platform recommends different products for them to offer their customers, based on the insights we see”, says Daniel.
“For example, based on usage patterns like a sudden short regular surge in power consumption we can recommend products like electric vehicle battery storage or suggest a better electricity plan for the customer. These recommendations help not only our retailer and utility customers, but also our customers’ customers. Everyone wins.”
At its core, Innowatts is an analytics and data product company. The company’s models have been trained on obfuscated data from millions of electricity meters from around the world and leverage these learnings to deliver accurate, AI-enhanced insights and knowledge, all while making its clients’ businesses processes more efficient and smarter. Innowatts’ platform helps energy providers accurately forecast load, design effective rate plans, manage risk, increase customer value and prepare for a sustainable future. Innowatts had helped its customers see a 40% improvement in forecast accuracy within 3 months, enhance customer lifetime value by $3,000 per customer, and avoid $4 million in Opex costs.
Druid has expanded far beyond its intended use case at Innowatts, which has led to expanding the insights that clients gain through the SaaS platform. Innowatts uses Druid to quickly work with massive data sets, and differentiates its platform with the intended direction to read meters and forecast usage on the fly.
“With the COVID-19 pandemic, there’s been a shift in electricity consumption from commercial to the residential side. Druid allows us to quickly provide BI innovation and sub-second querying to our clients”, says Daniel.
Daniel’s team runs regular batch ingestion on energy market data. This alleviates the overhead of operating an Apache Kafka cluster 24/7. Data arrives right before the market opens and the Innowatts platform creates a forecast immediately.
“Customers need to know where to buy power when the market opens. That’s when they want their reports. We ingest the most up-to-date information that we have. We know the exact time that we’re going to be producing the reports, so then we can ingest the freshest data into Druid. At that time, we spin up a cluster and use Amazon EMR, Apache Spark, and some other tools to prepare the data for ingestion.”
Daniel’s team scales their Druid cluster up and down every day.
“We know exactly what time we’re going to ingest data and how much. At that time, we increase the data servers on that cluster to speed up ingestion. Afterwards, we scale down because we don’t need such big machines.”
The team queries Druid via Plywood, a query tool created by Imply, because data is returned in a format that is easy to be fed into D3.js for visualization. Daniel says, “this workflow allows us to easily display actionable charts and graphs to our customers.”
“Customers are really happy with how fast our front end is, and that’s because it’s driven by sub-second query response times in Druid,” adds Daniel. He and his team find Druid’s documentation extremely valuable, as he says “it’s one of the best-documented tools we’ve seen, and that goes a long way with us.”
Daniel’s only been working with Druid for about five months, yet he and his team have built an industry-leading SaaS platform for the energy sector.