“How can we improve fan experience during a sports match?”
That was the central question for the 24-hour hackathon organized by Raccoons, Inspire Sports, and Pro League football club OH Leuven. During this hackathon, four groups of five people brainstormed ideas, prototyped demos, and presented their solutions to a jury.
To kick off the hackathon, our group started brainstorming around the following statement:
“If there’s one thing that defines a football match, it’s the crowd’s roar.”
After all, research has indicated that
- a team scores 0,45 more goals than their opponents
- a team has a higher probability of winning the match
- the referee would be more likely to decide in a team’s favor
… if their supporting crowd is larger and noisier.
Based on these findings, we decided to measure and stimulate ‘fan excitement’. Essentially, we wanted to engage and even incentivize the fans to be louder, more enthusiastic, and more supportive of their team, even (or especially) when the team performs less well than expected, or the match is less exciting for the fans in general.
Furthermore, we wanted to measure fan excitement on an individual level to incentivize fans personally. By working this way, we can feed these individual excitement scores to OH Leuven’s central data store to augment their knowledge about their fans.
Measuring excitement using biometric data
We used a Fitbit Versa 3 to measure fan excitement individually to collect the needed biometric data. More specifically, we used the following data from the (very easy-to-use and well-documented) Fitbit Device API:
- Heart rate
- Movement (from the accelerometer)
- Number of steps in the past minute
- Number of calories consumed in the past minute
This information was sent to the cloud every second. There, we created an algorithm that used these parameters to calculate an ‘excitement score’, a score between 0 and 100 indicating the level of excitement of a fan at a specific moment. The result was stored in a database and could be fetched by other systems (like Fitbit itself) using a RESTful API.
Engaging fans on a group level
To engage the fans on a group level, we wanted them to have a true impact on the stadium and game in a fun way. For example, we could use the different screens, lights, and speakers in the stadium to let the fans know they should act more excited. As a reward, the fans’ excitement is reflected through the screen, light shows, and music that plays increasingly louder.
When the (average) excitement of the fans fell below 50, the light strip would start to blink. When the excitement was between 50 and 100, the light strip would be brighter and brighter when the excitement reached 100. Finally, Neil Diamond’s “Sweet Caroline” (what else?) would play louder and louder when the excitement increased.
Engaging fans on an individual level
To engage fans on an individual level, we created an incentive. If they record their excitement during a match, they receive points in their OH Leuven app. In turn, fans could trade these points for rewards, such as a free shirt, free beer, or meet and greet with a player… We mocked this incentive in a mobile app. In this app, you can also see your current excitement in real-time. As a final feature, the Fitbit itself would start to buzz if the excitement score drops below 50.
Testing at OH Leuven - Antwerp
After more than 24 hours of hacking, preparing our demo and presentation, and finally presenting it to the jury, we watched the match OHL - Antwerp in the OH Leuven stadium. Of course, I wore the Fitbit Versa 3 and wanted to test my excitement score during the game. There were no lights or sounds in the stadium that I could influence with my excitement, but I could still check the score on the Fitbit and in the mobile app. To be honest, I was pretty impressed with the results.
At ‘dead moments’ during the match, my excitement score dropped below 50, and the Fitbit started to buzz (which did become a bit annoying at some points during the game). When I began cheering and moving, the buzzing would stop, and my excitement score would increase.
Of course, this is the first demo with a simple algorithm, and there are still many improvements to make. For example, we don’t use any audio information because the Fitbit Device API didn’t give us any access to the microphone of the Fitbit. Also, the algorithm was now manually tuned to my resting heart rate, and in general, it should be tuned dynamically based on the person wearing the Fitbit.
However, all in all, I believe the hackathon was a big success, and with some improvements, our project can improve the fan experience of a sports game. Unfortunately, in this case, the match ended with a 0-1 victory for Antwerp, which meant we didn’t get to hear “Sweet Caroline” one last time in the stadium.