How Urjanet Ensures Data Quality: An Interview with Urjanet’s Energy Data Quality Assurance Engineer

With a greater diversity and volume of energy data being generated and collected than ever before, the quality of that data becomes more critical to energy savings and achieving sustainability goals. You can’t rely on bad data to make important business decisions because errors can be costly. IBM estimates that bad data costs the United States over $3 trillion each year.

We often get asked about the quality of the data we deliver. Since our founding, we have processed over 16 million invoices from over 5,000 utilities in over 30 countries. It is safe to say the magnitude of the data we see on a monthly basis, and the scale at which we process that data, is unlike any other in the industry. Increased volumes of data often comes with increased complexity, and with increased complexity, mistakes, oversight, and errors can abound. This is why we take data quality very seriously here at Urjanet. Aggregating and processing our customers’ utility data is our core competency. Our corporate customers rely on quality data to run various energy management and sustainability programs, and our software and services partners depend on the data we provide to successfully run their businesses.

We interviewed our Energy Data Quality Assurance Engineer, Nikk Godha, to tackle the topic of data quality here in this blog post.

Q: Why is data quality important to Urjanet?

A: Data quality is based on accuracy and completeness, two of the main value propositions we offer our customers. In order to enable our customers to make smarter business decisions, we focus attention on making sure our data is both accurate and complete.

Q: How does Urjanet ensure data quality?

A: The data is checked at nearly every phase of its “journey” from the utility to our customers: from when it’s extracted from the utility, to when it’s stored in our database, and even while it’s being delivered to our customers. Our team tends to think of the data “journey” in those three phases: extraction, storage, and delivery. In the extraction phase, we create custom templates for every utility to ensure we can extract all the necessary data, so that the data set is rich and complete. With so many customers dispersed geographically, we’ve written over 3,800 utility integrations to support them. Next, during the storage phase, we have a comprehensive process to make certain the data we’ve collected is accurate. This includes nearly 150 audits that we’ve developed over our seven years in the industry, based on extensive domain expertise and widely-accepted best practices. These audits are like “checkpoints” that prevent bad data from being stored in our system. Customers can also customize and prioritize what audits are most important for their business. Then, we have the delivery phase, which is basically just the data seamlessly going straight to our customers’ applications. In this phase, we look at the data as if we were the customer. If our audits flag an error, then we bring a human element to reviewing the data. We poke holes through it, we’re critical of it, and we often play devil’s advocate to make sure we’re covering all aspects. This allows us to tailor the QA process to some degree for our customers’ specific needs.

Q: So what are some examples of the data checks or audits that we do?

A: We have what we call “Stage I” and “Stage II” data checks. Stage I checks primarily include data verifications that can be evaluated from within a single invoice. This verification process guarantees all data within an invoice is accurate. Some examples of Stage I checks include:

• Does the statement reflect a known commodity type?
• Are observed measurements and units known standards?
• Is the measurement period clear and can service dates be confirmed?

Stage II data checks are essentially a comparison of a new invoice against a “history” of previous invoices to confirm that cost and consumption fall within rational historical ranges. Stage II data checks can often test for seasonal variability and more subtle pricing changes over time. Some examples include:

• How does an invoice compare to other invoices from the same utility provider?
• How do the charges compare to other invoices with similar rate plans?

Q: What’s the most difficult part of achieving high data quality?

A: Some of our customers have pretty strict requirements around turnaround time. They usually need the data delivered pretty quickly. Managing data quality alongside speed of delivery can sometimes be challenging, especially in situations when our audits signal errors. When that happens, we have to look at what happened, for example, if there was a format change or new data point on the invoice.

Q: What do you think is key to our success in delivering high quality data?

A: Well, our seven years in the industry has given us tremendous experience in seeing all the different kinds of scenarios and situations that a utility provider can throw at us. There are more than 5,000 utilities we’ve acquired data from. Our audits have seen so much data over the years that they know the nuances of particular utilities. It’s self-learning. And now that we’ve gone international for almost two years, we are also becoming experts in delivering high-quality data in multiple languages. The breadth and depth of knowledge we have on how utilities present bill data is unparalleled in the industry.

Q: What’s next for the Urjanet Data QA team?

A: We’re constantly adding and evolving the data checks in our audit library as we take on more customers, especially for those with fairly complex billing scenarios. The more data we run through our platform, the smarter our audits get. We’re also turning increased attention towards fortifying quality assurance on the delivery side so that we have that extra layer, allowing us to look at the data as if we were the customer. We’ve heard good feedback from our customers so far, so we only hope to continue growing from here.

To learn more about our data quality assurance process, feel free to contact us.

Related Resources:

• Webinar: 5 Ways to Ensure Data Quality for Sustainability Reporting
• Urjanet’s Guide to Utility Tariffs
• Big Energy Data 101

---

---