IBM expanded its three-decade partnership with Wimbledon this year to create a new digital tool to encourage fan engagement. The tool collects insights into every player and every game and provides users with new data about the participants in the tournament and hopes in the longer term to stimulate interest in players beyond just the Wimbledon season.
IoT World Today took a look inside the control room where all these insights are collected, and got a little more information about how it works.
“We’ve built an innovation platform that is really about turning massive amounts of data into meaningful and compelling insights,” said Kevin Farrar, IBM’s UK sports partnership leader. “We generate data in real time, stroke by stroke, and combine all of these tennis stats with other data sources to create a holistic understanding of a player.”
These other data sources include media reactions to a player or game, which are combined with the player’s statistics to make a prediction about how likely they are to win and what their career path has been so far. This data is then fed into the IBM-designed platform to generate digestible insights.
“We have a different variety and volume of data”, said Tyler Sidell, technical program manager of sports and entertainment partnerships at IBM. †We have unstructured or structured data, which you can see coming in on the left [of the diagram]† It’s all because of the system to create the experiences on the right – to Wimbledon.com, IBM slam tracker, power index, match insights. What you see in the middle is all the magic. We have a hybrid cloud architecture, so it’s made up of private and public clouds, so we can write one and use it anywhere.”
“We are able to actually automate the process of creating all the insights for the match,” adds Sidell. “We use the explainability of AI as a toolkit to create a player profile.”
While data capture has been a part of Wimbledon to some extent, the richness and accuracy of data collection tools has naturally become more defined and the club has been able to adapt to consumer demands.
“Part of the feedback from the club last year was that they wanted bite-sized content as different fans consume information in different ways,” said Farrar. “You see some fans consuming stuff on social media and mainly interested in videos or short content, while others want a high level of detail. We try to serve all those different audiences, so what we have given is a high level summary of the competitions you can choose to dig deeper into, and adjust accordingly who receives what content “An important thing this year is also explainability; it’s all explainable AI, so it gives fans confidence in the technology.”
The AI platform creates match insights, can predict who is likely to win, and can even generate various highlights from a game using audio and visual data from both the crowd and player reactions. However, a team of editors must approve all these creations before they are published, and Farrar stressed the importance of preserving the human role in the system.
“It’s about increasing the human aspect of it, so it’s not AI replacing humans, it’s about increasing and assisting them,” Farrar said.