Smart Meter Readings
OwlDQ discovers billing errors in smart meter data. To date Owl has processed over 43 million meter readings and over 7 years of billing data to unveil historical trends and data points that don’t add up.
Owl’s initial scan was insightful and revealed simple issues in our data. However once we pivoted the Machine Learning to focus on YoM (years over months) it got even smarter. It created training sets that stacked December over December for the last 5 years instead of creating training sets in a standard time-series such as Dec, Nov, Oct, Sept. This revealed hidden issues and relationships in our data.
Owl's replay feature scanned several years of data automatically and within 20 mins of having it installed we could focus on the most expensive issues to our business.