Amazon Web Services (AWS) is the world’s largest cloud provider. AWS’s primary business is providing cloud infrastructure for your applications and data. However, AWS does much more, and one of the things that is less well-known is AWS for Sports. The AWS for Sports offering contains various solutions through which AWS helps sports clubs perform better by leveraging data. AWS is doing this in several sports.
AWS is active in Formula 1, Bundesliga (German soccer league), NFL (American football), NHL (American ice hockey) and PGA Tour (golf). It also cooperates directly with all kinds of sports clubs from various sports. The technology can make sports teams perform better or become more attractive to viewers. By using AWS for Sports, a club can make better decisions based on data to improve the team’s performance.
We talked about it with Luuk Figdor, Principal Sports Advisor at AWS. In this article, we focus mainly on improving performance within a sport.
Also read: AWS introduces Glue Data Quality for data quality
Improving sports performance with data, Moneyball is more real than ever
For improving a club’s performance, more and more data is available. Generating that data requires many sensors. In soccer, the camera is one of the most important sensors. Soccer clubs that are actively engaged in data often have 16 to 20 cameras hanging in their stadiums to continuously follow the players on the field. Their position is recorded 25 to as many as 100 times per second. Each movement is analyzed through algorithms.
Another way to collect data is to have players wear a sensor. At soccer matches, you sometimes see a player take off his shirt and wear a kind of sports bra underneath. They wear those because there are sensors in them. This measures what a player is doing, what movements they are making, how intensive they are performing and how that performance declines during a game. It offers insights that coaches didn’t have before.
One of the most popular data analytics within soccer is called skeleton tracking. This involves using sensors to track a player’s skeleton down to milliseconds. With this, a player’s every movement can be tracked. Walking, sprinting, running, jumping, shooting, passing, and receiving the ball with the left or the right foot; in fact, every movement of a player becomes visible in the data. The intensity can also be measured with this, making the so-called skeleton measurement extremely valuable.
AWS builds an ecosystem of partners to generate sports data
As you might imagine, a vast range of expertise is involved in developing sensors, as well as the algorithms and AI models needed to properly read, interpret and convert those sensors into usable data. AWS does not do this itself.
In recent years, the company has gathered several partners who are good at developing sensors and models that can provide valuable data from various sports.
AWS helps clubs connect all the data in a Sports Data Platform
The knowledge and expertise of AWS comes into the picture when a sports club has several active sensors generating data. That’s also where the most significant challenge begins for many clubs. Purchasing, installing or using sensors is the first step. A sports club then has all these silos of data that are relevant on their own, but really valuable if you can combine different sensors.
That’s where a Sports Data Platform comes in. This solution from AWS can ingest data in all sorts of ways. From S3 buckets (object storage) to live streaming data. It can tie all those different data sources together. If, for example, data analysis of skeleton tracking shows that a player often gives too much and overloads himself at specific points in the match, the combination of data gives the ability to visualize this by showing the video at those specific points. The conclusion may be that the player should try to make fewer slidings or jump less unnecessarily. Or it could be something technical, the way the foot is placed when starting a sprint. By visualizing the problem, the coach or physiotherapist can better analyze it and communicate it more clearly to the player.
According to Figdor, the Sports Data Platform is at the heart of sports innovation. Because clubs can do much better analysis on such a large amount of combined sports data. They get the ability to extract more value from it. What started with simple analytics has now grown into advanced analytics based on AI and machine learning. The more data you have about a player, the easier it becomes to get better analysis and improve player skills, but also to manage a player that he doesn’t do too much.
In many dugouts, nowadays, sits an assistant coach with an iPad watching live analysis of data coming from a Sports Data Platform. More and more big clubs are employing their own data analysts to extract value from the data collected. Als more companies are emerging that specialize in this field that develop sports technology. They can provide AI models based on data to do analysis. Ultimately, it’s a whole new specialized industry.
How does a club start with data analytics and for which clubs is this feasible?
Sports clubs currently not doing much with data may find the approach interesting. At the same time, the question may be how to set this up in their club. Is it also feasible for the club?
Surprisingly, it is feasible for more clubs than they sometimes realize. Every club now works with data in the Bundesliga, partly because the league itself invested in this. As a league, it wanted to give the clubs a technological edge over other European clubs.
Figdor indicates that some clubs sometimes see it as something new, so they have to find new budgets within the organisation. This is even though they often already have several data tools internally that are all a piece of the same puzzle. For example, physiotherapists and doctors at a club already work with digital tools to measure players’ performance and intensity at games and training sessions. A club’s scouting department often works with video footage to do analysis. Trainers and coaches also often use video analysis to improve plays and strategies. Sometimes data is also sourced externally based on video footage. How many passes did players make, how many shots did they make a how many steps did they take?
However, all these data tools are bought separately. Within clubs, the different departments sometimes don’t even know from each other what data they have or how they could use it. Some solutions are easy to use in an AWS Sports Data Platform, while others are not. But by making these investments at a higher level within the club, all stakeholders in the club can have better data and analytics.
In addition, Figdor argues that this investment should be separate from club politics. Typically, each trainer brings his own technical staff. If clubs also go along with that in terms of data solutions then it does become an expensive proposition. If you have to implement an entirely new data platform when a new coach is installed, you are talking a complete replacement every few years, or at some clubs every year. Figdor believes that if you have a well-functioning data platform with good usable analysis, every trainer should be able to use it.
More and more clubs are investing in data
A few years ago, this form of data analysis in sports was only feasible for the really big clubs. Meanwhile, there have been considerable developments in AI, machine learning and the availability of sensors. The AWS sports platform has also gotten better and more effective over the years. It makes the new industry more extensive, more competitive and also more affordable. This allows more and more clubs to adopt a data-driven approach. The familiar top-down approach of AWS can also be seen in sports data.
Figdor argues that in addition to the overall cost, there are many degrees and choices in which you can work with data. You can analyze a player 100 times per second or just 25 times. Are you going to work with live footage which is much more costly or not? What sensors and AI models are you going to use for data analysis? As a club, you can go for everything immediately and make a huge investment, but most clubs start small and then build out the platform incrementally. With an active platform, the internal data analysts at the clubs can also more clearly identify what data or sensors they need to achieve better results.
Ultimately, Figdor hopes he can connect more and more clubs that will build their Sports Data Platform on AWS so that AWS can play a central role in their data strategy. In addition, he hopes to find even more partners. That way, the AWS Marketplace can play a more significant role and boost the sports technology industry. The marketplace should give clubs a destination to buy more AI and machine learning models to unleash on their data for better analytics.
We know that at AWS will continue to drive innovation in sports with its sports solutions.
17 hours ago
