The growing electric vehicle (EV) market has a highly decentralized approach to charging. Drivers may plug-in their car in their garage or use stations, operated by networks such as Tesla, ChargePoint, EVgo and Blink, found in a wide variety of locations.
A significant constraint on the growth of the EV market is the absence of reliable and accessible data on charging prices. Addressing these data issues, according to Dr. Tim Coburn, professor, data analytics, and interim chief data scientist, can help shape charging behaviors.
“Access to a comprehensive source of EV charging price data can facilitate decision-making for EV operators, charging station operators, policymakers, and business innovators. However, such data do not yet exist in a clear and accessible format,” says Dr. Coburn.
Much of EV charging data is proprietary and difficult to access. But Dr. Coburn and colleagues David Trinko, systems engineering Ph.D. student at Colorado State; Thomas Bradley, chair, Systems Engineering Department, Colorado State; Emily Porter, senior associate, Rocky Mountain Institute; and Jamie Dunckley, senior project manager, EPRI; were able to explore a dataset from the crowd-sourced platform PlugShare.
PlugShare has EV charging data platform with price information and other metadata for a substantial portion of stations in the U.S. The authors applied an ad hoc text mining approach because PlugShare’s data is in a semi-structured textual form. That method included a combination of human intervention, original scripting and machine learning to extract quantitative price information into a new dataset.
They then applied descriptive analytics, in the forms of graphs, and interpreted data to summarize the quantitative data. Those analytics provide a high-level overview of public EV charging prices and reveal a wide variety of prices according to location, the network operator, and type of location.
The authors cite that while this dataset is informative for both consumers and operators, it is one that consumers cannot easily access. The crowd-source nature of the data also makes it difficult to assess its quality.
“However,” says Dr. Coburn, “this research is one of the first efforts, if not the first attempt, to actually explore real data for EV charging.”
Dr. Coburn believes that further research on EV charging could improve EV adoption and facilitate infrastructure expansion.
“Absent the creation of an improved, standardized, and more fully accessible dataset, both consumers and those considering developing a business around EV charging are somewhat limited in terms of actionable knowledge,” says Dr. Coburn.
Such a dataset could have a significant impact on the EV market.
“Of course, we really don’t know too much about EV consumer behavior right now,” says Dr. Coburn. “The market is growing quickly and that could just be due to an environmentally-conscious bandwagon effect.”
However, it is also possible that improved data will help operators make better business decisions as charging infrastructure expands. It could also lead to further EV adoption by consumers.
“More accessible data could help consumers know where to charge, which stations offer better prices, and how vehicle charging would affect a long-distance trip. Processing this information also could encourage more people to purchase a car,” says Dr. Coburn.
Pursuing these issues could lead to an impact that goes far beyond the expansion of the EV market. It could ultimately help offset climate change by reducing dependency on fossil-fuel transportation.
“All of these issues are important because we are in the middle of an effort to try to entirely change society and its fossil-fueled economy,” says Dr. Coburn. “It’s going to take a lot of work to totally displace our reliance on gasoline- and diesel-fueled vehicles and alter our petroleum-based way of life.”
The full paper, Combining Ad Hoc Text Mining and Descriptive Analytics to Investigate Public EV Charging Prices in the United States, is available online.
