Subscribe to the SportTechie Podcast
TuneIn | iTunes | Soundcloud | Stitcher
Alex Turnbull, director of broadcast production at the PGA Tour, and Mike Downey, principal evangelist for sports and entertainment at Microsoft, joined Bram on the SportTechie podcast to talk about golf’s new content relevancy engine. The CRE was built by Microsoft for the PGA Tour, and uses artificial intelligence to help broadcasters automatically find statistics and content to enhance their coverage.
“Essentially there is an opportunity for us to take the burden off of both our employees and our broadcasters covering the sport. And that’s where we began working with Microsoft,” Turnbull said. “And we’re looking for ways where we could apply automation, apply logic, apply artificial intelligence to surface relevant statistical information and present that to our fans and our viewers and players to make an otherwise mundane situation dramatic and compelling. So that’s where we began working with Mike and his team to create the content relevancy engine.”
Alex Turnbull: Broadcasters might not be able to tell man from machine
“The beauty of this tool is that a lot of times they might not even know the machine is providing them with this information, because they’re just getting the information. Little do they know what’s actually happening in the background, or who it’s coming from or where it’s coming from. They’re receiving it through some of the tools that we have, or they’re receiving it graphically on air. And they may still think that the team’s back there working on that when in reality it’s actually the content relevancy engine surfacing that data.”
Mike Downey: How to build to use data to enhance the golf broadcast
“It’s a tough challenge for a bunch of data scientists here at Microsoft, but luckily most of us are big golf fans and we have personal incentive to figure this out. What we really did is look at how can we use things like machine learning, predictive analytics, the ability to really go in and say ‘Alright we have a huge data set here to work with. Let’s find the patterns, let’s devise tools that learn over time, so that we can score things and score results, we can learn from trends of what data’s being used and what data’s not being used.’”
Alex Turnbull: Fans can use second screens to follow their favorite players
“We had a very small team, and we could focus on the broadcast with our small team, but we couldn’t focus on every single player in the field. The CRE brought scalability to that, where we can now supply a steady stream of relevant information for every single player, for every shot in the field. And therefore allowing fans to follow any player that they want on a second screen or on a smartphone or iPad, and follow along online digitally, and get compelling content while they’re following their favorite players.”
Mike Downey: How to enhance content from before the age of data
“We can find patterns in the images themselves. We can figure out ‘OK, so that’s Tiger Woods in 1997.’ We can figure that out, go through all of the footage, and find all of the clips. We can look at the broadcast transcript and use that to correlate the video clips with the time codes. We can do optical character recognition on the video so when the broadcaster put up the onscreen graphics we could actually go back and read those and interpret them as text. And that can go back as far as we want, depending on how high quality the footage itself is.”
Hear the full interview with Turnbull and Downey, including their thoughts on how they might improve the content relevancy engine, and how players might be able to leverage its insights, on this episode of the SportTechie Podcast with Bram Weinstein.
Contact us to suggest a future guest for the show.