Rajiv Maheswaran and Yu-Han Chang with Second Spectrum CTO Jeff Su and students
As the world has become increasingly complex, “Big Data” has become a phrase that has permeated every industry. However, while you expect to hear it in telecommunications and finance, over the last year, it’s found a new home: sports.
This new field of sports data analytics is the basis of Second Spectrum, a company founded by USC Viterbi Research Assistant Professors Rajiv Maheswaran and Yu-Han Chang that counts the Los Angeles Clippers among clients. The duo has also developed a yearlong sports data analytics Capstone class that was made available to computer science majors for the first time this school year.
“In the last year, teams in all major sports have accumulated massive amounts of tracking data, and they’re starting to investigate it,” said Maheswaran, who co-teaches the class with Chang. “The amount of information available to them has increased by three to four orders of magnitude, so a single statistician can’t get it done anymore. It’s become an engineering-level problem.”
Beginning to address this engineering-level problem is a handpicked team of six students, who spent the first semester learning a programming language with which they will build a better sports video-viewing platform. The student group is working on a system for both machines and humans to index and annotate video. Viewers would then be able to re-watch key moments of a game based on things like the excitement of the crowd, the play-by play feed or reactions on social media. It has the potential to fundamentally transform how people watch video and in particular, sports.
Response from the student team has been overwhelmingly positive.
“I was originally enrolled in the computer science Capstone class like everyone else, but from a structure standpoint, I would have spent the entire first semester not really being able to do anything other than conceptualize,” computer science major Jeremy Chen said. “So when I got an e-mail from my advisor for an alternate research class that would enable me to delve right into sports data analytics, I jumped on it.”
The students have chosen to focus on soccer video initially, but they have a larger vision: to make their code general enough to be applied to all other types of video. A soccer game can be indexed by anything from the clock to the crowd noise, but an alternate application of the same technology would also enable movie fans to isolate scenes in a movie by a famous quotes or the type of background music.
“We want to develop a system in which we can ultimately categorize any kind of video by anything,” Chen said. “We’ll spend the next semester trying to figure out what is possible and what is useful."
Chen, a senior who already has a job offer from Expedia in San Francisco, actually began his studies at USC as an aspiring sports journalist before changing majors to computer science. However, once he took the rather unorthodox leap from The Annenberg School for Communication and Journalism to The USC Viterbi School of Engineering before his sophomore year, he realized that a career as a software engineer was incredibly versatile.
“One of the cool things about computer science and software in general is the fact that you can go literally any direction with it,” he said. “You could do sports or travel. My first internship in school was with the Southern California Earthquake center. You can do anything.”
Despite the fact that Maheswaran's and Chang's company and class are centered on sports, they encourage open-ended thinking like Chen's.
“The idea of this class is to give the students direct access to elite industry professionals,” Maheswaran said. “The end result of their project is secondary to us teaching them a way of thinking and letting them run with it because that’s how things work out in the field.”