Faculty, Students Using Big Data to Analyze Energy-Use Patterns
A team from the School of Information Studies (iSchool) is conducting research analysis using big data sets from the Pecan Street Research Consortium, a global collaboration working on utility system operations, climate change, integration of distributed energy and storage, and customer needs and preferences.
Consisting of graduate students under the guidance of Associate Professor Jason Dedrick, Professor Jeffrey Stanton and Associate Professor Murali Venkatesh, the team has joined a number of other research universities also involved in analyzing how consumers use home energy and how their usage patterns impact the power grid. The research ultimately has the potential to launch industry-wide changes in the way consumers use and pay for energy, how utilities plan for peak-use issues, and how the electrical grid system can be optimized, according to Dedrick.
Using huge data sets provided as open-source information by the consortium, the team is working with millions of time-stamped electricity records from the Pecan Street Research Institute’s original field research. Researchers can analyze data taken from meters that have collected utility-use data on disaggregated electricity use from 50 households at 1-minute and 15-minute intervals daily, and from rooftop solar panel generation, electric vehicle charging, residential transformer loading and original data from pricing trials and demand response and default setting behavioral trials.
The iSchool is able to participate in research and analysis of data sets that large because it has the capacity of on-site IBM System z mainframe computers, according to Dedrick. At present, iSchool students who are working on the project are converting the data to the z10 and z196 mainframes, and determining how to present it, he says.
“The question is how to put it in a format that can be communicated and useful. We’re working through that and thinking what kinds of analyses we can work on. We’ve got a lot of ideas already; it’s really kind of a gold mine for data,” Dedrick says. Among the kinds of studies that could be generated from the data sets are ones that may have potential for funding from the National Science Foundation or other sources, Dedrick says.
The kind of detailed analysis that is possible with huge data sets may one day contribute to the perspective of energy consumers as well as energy pricing and policy by illustrating patterns of use that can incentivize conservation, Dedrick says.
“One of the things where we’d really get a lot of value out of this is if electricity was priced at its actual cost. Because the true cost of producing electricity at different times is currently invisible to people, they have little incentive to save energy. But if you knew that the price is 50 cents now versus 5 cents later this evening, people would really think more carefully with how they use electricity,” he says. Ultimately, though, the findings may be more useful to utility companies themselves, Dedrick believes, since the information derived may inform and illuminate planning for capacity, peak use and grid optimization.
“That’s where the value of big data comes in,” Dedrick says. “We’ve got the hardware, we’ve got the big data sets, we’ve got the tools we’re developing, and that’s going to give us a big selling point for the school.”
With the capacity for data analysis that the mainframes provide, Dedrick also envisions the iSchool becoming a place “where if the utilities have a problem to solve or want to do some analysis, but don’t have the time or resources to do it, we could be kind of an R&D shop to run some of that kind of data and develop models.”
“As far as students go, this is a big career opportunity, since there’s a large demand for science skills and the ability to work with large data sets. To have a place where they can get hands-on experience like this is going to be really good for a lot of students,” he says.