CACHE Affiliates Dive Into New Chicago Electricity Data


After sifting through 12 months’ worth of northern Illinois electricity use data, Illinois scientists are eager to dig into the patterns that may reveal the human side of energy use trends.


In most areas of the world, electricity is a fundamental part of daily life. It heats homes, cycles clean air through buildings, and lights up almost everything we see. But despite its universality, we actually know very little about how people consume electricity in their homes.

For scientists who study energy systems, residential energy efficiency, and the health impacts of indoor spaces, the methods for learning about people’s electricity use are limited. Behavior surveys and utility bills can show consumption trends, but aren’t feasible on scales larger than a neighborhood or two.

To reveal broad behavior trends that drive residential electricity use, researchers at the University of Illinois are turning to smart-meter electricity data from the ComEd utility in northern Illinois.

The data were obtained through a partnership with Environmental Defense Fund in early 2018, under a new program that allows ComEd to release anonymous electricity usage, in 30-minute intervals, for most in ComEd’s northern Illinois territory, including all of Chicago.

It’s a huge amount of highly-specific data, and before researchers can hope to draw conclusions from it, they need to understand the landscape the data presents. Ashlynn Stillwell, Professor of Civil and Environmental Engineering at Illinois, Paul Francisco, Senior Research Engineer at the Illinois Applied Research Institute, and Lucas Djehdian, a Civil Engineering Master’s student, took on the challenge.

Funded by the Center for Applied Collaboration on Human Environments (CACHE), Djehdian spent three months combing through the ComEd data, arranging them in an easy-to-use format and drawing some preliminary trend charts.

“There’s a lot of things we could do with these data. At this point, we’re doing a lot of exploratory analysis to get familiar with them and start to brainstorm what the most interesting questions are,” said Stillwell. “The thing about data science is that a lot of time you need to see what you’re working with.”

To create the first pictures of electricity consumption, Djehdian sorted buildings by ZIP code and color coded a map of them by average monthly level of electricity use. It’s an instant representation of high and low energy users. What makes some areas more electricity-intensive than others? That’s the kind of question researchers are planning to ask next.

“Since we have data points every 30 minutes, we can see trends over a day, a month, a season, a year. We can zoom in and out of both geographic area and time. That’s a lot of options!” said Djehdian.

“Just taking time to put the information all in order, and then letting other scientists see the initial charts, I think that’s a good step to move on to more specific studies,” he said.

“Based on Lucas’ high level view, we can dig into what’s going on. There’s a lot we can learn about behavior by not interacting with the humans themselves but looking at the numbers.” said Stillwell.

“Should we focus on single family homes with electric space heat and start to look at patterns in December, January, and February because those are the months that jump out of us? In this big sea of data, where do we start to dig deeper to start to visualize things? That’s what this project has been all about.”

Posted in News