Written by Jessica Schuler
As the Cleveland Indians and Toronto Blue Jays meet in Major League Baseball’s American League Championship Series this week with a trip to the World Series on the line, Blue Jays’ slugger José Bautista’s bat has been quiet. Perhaps the Indians have sought out K12’s big data team for tips on how to get perennial All-Star out.
K12’s big data team – which consists of senior director of systems and technology Adam Paul, senior director of product management Ben Graff and data scientist Haitao Du – won the 2016 Baseball Analytics Hackathon presented by the Baltimore Orioles and consulting firm Booz Allen Hamilton by answering that very question, which many MLB teams have struggled to do for the last seven years.
At K12, the big data team focuses on analytics and adaptivity, predicting at-risk students from an engagement and academics perspective and devising differentiated paths for students depending on their proficiency within the curriculum. The team looks for learning patterns and effectiveness of learning solutions to help inform future instruction and curriculum and program development.
When Adam learned of the event from a friend, he thought it would be a fun way to see how K12’s relatively new big data team compared to others in the field.
“We thought it would be a great way to see what other folks in the area are capable of and what they do,” he said. “As a new team, we thought it would be a good bonding experience and would help sharpen our skills and let us know how we were performing against some of the best companies and best universities in the space.”
The 21 participating teams were provided with a large data set called PITCHf/x for every pitch that has been thrown in a Major League Baseball game over the last three years. The teams were given an open-ended mission: use the data to solve a problem that is actionable in the game and present it back to the judges in eight hours.
“I think the most challenging part was the problem definition,” said Haitao, who hails from China and had never seen a baseball game prior to the competition.
After realizing their original problem was too complex for the time they were given, the K12 team shifted their focus to determining the best sequence of pitches to get a given hitter out.
“We went through all the data and said if you were going to face this hitter, here’s the different pitch sequences and locations that you could use that would be most effective to least effective at getting him out in a given situation,” Adam explained.
While the trio calculated for every hitter in baseball, their three-minute presentation focused on the Blue Jays’ Bautista, whom Ben described as the Orioles’ “arch nemesis.”
The judging was based on what the teams did with the data and how well they presented the problem and its solution.
“If you did really fancy math and solved a really hard problem but no one understood what you did, that didn’t go so well,” Adam said.
The K12 team won the grand prize – tickets to watch a 2016 Orioles game in a luxury suite.
“Essentially what it shows you is the level of competence that we have in this area,” said K12’s chief technology officer and senior vice president of software product development, Bala Balachander, who oversees the big data team. “We didn’t have a big data team three years ago. We knew we had a good team but it is great to get validation from others. Big data analytics is a very strategic area for K12 and I think we are positioned well with this team.”
Bala credits the composition of the team for the big win – “Adam is really solid in engineering; Ben is a product person, so he’s thinking of it from a customer perspective; and then you have the brains of the data scientist in Haitao,” he said. “The three of them are extremely good individually, but more importantly they are really good from a team perspective. They make a great team, and I think that was one of the contributing factors to the win.”
When the trio returned to Oriole Park at Camden Yards this summer to attend the game with the tickets they won, Adam asked a member of the Orioles’ baseball analytics group if they had used their solution to get Bautista out this season.
“He smiled and politely said they couldn’t share that information,” Adam said.
However, one look at Bautista’s 2016 stats against the Orioles may provide the answer. His numbers against Baltimore were down dramatically from the previous season – from a .373 batting average with six home runs and 20 RBIs in 2015 to a .204 average with three homers and five RBIs this year.
Maybe the K12 team was onto something!
From left: Haitao, Lucas (a friend), Ben and Adam in the luxury suite at Camden Yards