Professor Zoran Obradovic becomes animated when talking about “big data.”
“I’m working in situations where data is collected over time and space,” he said, before displaying several complicated graphs.
Obradovic, the director of Temple’s Center for Data Analytics and Biomedical Informatics, analyzes graphs with millions of data points to predict future events. The topics range from the price of medical care to the connections made on social media.
“I try to build predictive models from data,” Obradovic said. “We observe and then we try to guess the future to the extent possible.”
He recently joined an interdisciplinary team that will be evaluating a massive amount of statistics about digital activity. The group, which includes Obradovic, management information systems professor Youngjin Yoo, biology professor Rob Kulathinal, and management information systems professor Sunil Wattal, received a $900,000 grant to be used over three years from the National Science Foundation.
The NSF has given grants to the project, led by Yoo, in four of the past five years.
Yoo said the team will focus on two subjects. First, they will look at GitHub, a website that allows users to host and share open-source software. The researchers are interested in how the programmers who use the service interact.
“We want to see how different patterns affect performance,” Yoo said. “When a programmer is changed, how does it change the relationship?”
The team also plans to study the connections people make on Twitter, looking primarily at how messages are communicated.
The goal of the investigation is to predict behavior based on the millions of traces made by users of the websites.
“If we can find patterns, we should be able to predict,” Kulathinal said.
Kulathinal, who has worked with Yoo for several years, said the work is “cutting edge” and he is looking forward to collaborating with the other researchers.
“We’ve already recruited the best people in business, biology and computer science,” Kulathinal said.
“It’s exciting to learn new things,” Yoo added. “We can transcend the boundaries of the different disciplines.”
Obradovic, who joined the group for this project, was attracted to the study by the academic diversity present in the team.
“What was to me appealing is that this is a very interdisciplinary team,” he said. “This kind of work can be done better when you are in that kind of group.”
If not for the $900,000 NSF grant, the team may never have come together, Kulathinal said.
“The grant allows us access to Obradovic’s lab,” Kulathinal said.
The grant facilitated the connection between the researchers and set the study in motion, which they had been discussing for about a year, Yoo said.
“The grant makes it easier to collaborate,” Yoo said. “It forces us to do something because there’s a deadline.”
The four professors, who have varying interests, chose to focus on digital technology because of the size of the data that has been collected. It is their job to take facts from data and find patterns and processes. Often, the statistics are recorded before the researchers know how to handle the information.
Interpreting vast quantities of data is not easy, they said. “Big data,” which has become a buzzword in the business world, poses new difficulties in Yoo’s view.
“It’s a tremendous challenge for computation,” he said. “We’ve always had data in volume, but this data is granular, unstructured trace data.”
Once the data is gathered and analyzed, the researchers hope they will be able to predict certain behaviors and the effects of those behaviors.
Jack Tomczuk can be reached at email@example.com.