There is an entire industry built up around deciphering where 16- and 17-year-olds will play college football. Websites boast “crystal ball” predictions of where top high school recruits will suit up. Companies charge for premium subscriptions with claims that they can decode the caprice and whimsy of children.
She developed a mathematical model that predicts with 70 percent accuracy where a high school football player will go to college. And she uses nothing but their basic biographical information and Twitter account.
In other words, she can read the minds of some of sports’ most sought-after prospects by reading their tweets and looking up some basic biographical information. Her paper on those findings was published this month in the INFORMS journal “Decision Analysis.”
In other words, she could completely own Nick Saban and Urban Meyer on the recruiting circuit if she really wanted to, with just the use of a fancy spreadsheet and some half-decent computer code.
“If you really want to see where someone is committing, you shouldn’t overlook [social media] data,” she said.
Bigsby developed the model as part of her PhD program in information science. She wanted to study wrestling recruiting — Iowa consistently fields a top wrestling team — but went with football because the sport was more popular and because of the national obsession around recruiting classes.
She mined data from 573 athletes in 2016 from the 247Sports recruiting database who had at least two Division I scholarship offers and public Twitter accounts. Then she pulled their tweets, followers and accounts they followed each month and distilled the data into a model that makes it all easy to understand.
She found that if a recruit tweeted a hashtag about a school, his likelihood of committing there jumped 300 percent. For every coach the athlete followed from a given school, his likelihood of committing went up 47 percent. When a coach follows an athlete, likelihood increases 40 percent.
“The most significant actions online are the actions the athlete is doing,” Bigsby said. “Who is he following? What is he tweeting? What hashtags is he using?”
Her model crunched those numbers along with other data sets — i.e.: a college’s location relative to the recruit’s home town, a college’s academic ranking, a college’s recent football performance, and more — and spit out a list of universities a recruit was likely to attend, along with each school’s odds.
The model correctly predicted a recruit’s choice 70 percent of the time. And if the model was wrong, recruits generally chose the “second-place” college, Bigsby’s paper shows.
“We can narrow most people’s choices down to two schools,” she said, “but you never know what teenagers are thinking.”
The model could provide better predictions, Bigsby said, if researchers pulled recruits’ Twitter data every week instead of every month. Plus, she’s still tweaking the model to better interpret what tweets mean.
An athlete posting “Just got an offer from Iowa,” and “Can’t wait to visit Iowa,” mean very different things, Bigsby said. The first is self-promoting, and probably doesn’t do much for the Hawkeyes’ chances of landing a commitment. The second one is “ingratiating.” The athlete is trying to join an online community conversation about Iowa. That certainly helps the Hawkeyes’ odds.
So imagine the following: Alabama beats South Carolina and then has a bye week. Crimson Tide assistant coaches fan out across the continental United States on recruiting trips equipped with weekly reports on prospects’ online activity and their current likelihood of choosing Alabama.
That’s what this model can do, Bigsby said. It can really give teams an edge in the valuable, year-round recruiting game.
Only one problem: You need an information science expert to run the numbers. Bigsby has a potential solution for that, too.
“I’m careening toward graduation,” she said. “If a football team wants to call me, I will certainly pick up the phone.”
More from The Post: