You are here

Google Research Grant to allow Ohio State to analyze live streaming data

Assistant Professor Yuejie ChiA grant to the Department of Electrical and Computer Engineering from the Google Research Awards program will allow Ohio State researchers to study ways to improve analysis of massive amounts of streaming data, without actually storing the whole data.

The resulting data analysis tool has potential applications in everything from social media analytics and national defense to predicting behavior in financial markets.

Yuejie Chi, assistant professor of electrical and computer engineering, said there are countless examples of vast amounts of streaming data, ranging from click counts on websites to stock market prices, from smart utility meter information on a power grid to tracking moving targets for the military.

She said collecting the data is relatively simple. The challenge is in making sense of that constant flow of information and making predictions based on those calculations in real time, with minimal cost, minimal digital storage and in ways that can produce beneficial actions.

“It’s about how we extract the information,” said Chi, who received the one-year, $44,969 grant in September for her proposal. “The appealing aspect of this approach is that we don’t have to store all the data and we only need to scan it once.”

Chi said it is possible to use mathematical algorithms to make “sketches” of live, streaming data and extract what is valuable. For other approaches to big data analysis, digital memory cost can be high because of the need to access original stored data. An algorithm that performs its calculations on live streaming data minimizes that issue, she said.  

The potential applications for Chi’s recommended approach are numerous for anyone who has large amounts of constantly flowing information and a desire to understand and apply it. For example, companies like Google get massive amounts of information 24 hours a day from everyone using their service on the internet. Without the ability to quickly determine what that data means there is no way to apply it to commercial, research or other practical applications, which Chi said could be innumerable.

The grant for her proposal, “Covariance Sketching of High-dimensional Stream Data,” will fund a stipend for a graduate student to work with Chi on the research.