Summer means football and right now we get to enjoy the FIFA Women’s World Cup in France. I have personally been really enjoying watching the women’s games and the different dynamics on the pitch – as well as the enthusiasm surrounding the game on social media and among fans.
It’s refreshing to see football from a different angle while still enjoying a sport that is a favorite with hundreds of millions of people around the globe.
Behind the high-octane performance we watch for 90 minutes there are years of training, commitment and passion driving players to be their best. They’re surrounded by teams of experts helping them take their physical and mental performance to the next level and developing team tactics. And behind these people there are further experts, many of them using data to identify opportunities, trends, spot outliers, find insights and develop a deep understanding of opponents and their playing styles.
Last week, at our Exasol Xperience 2019 conference we brought together 400 customers, partners, users and some of our own people for discussions, talks and seminars on taking analytics and data further in organizations and accelerating businesses well beyond their existing data frontier.
During the event we hosted a panel discussion about the use of data in football. There are many parallels between a football club and a business in more traditional industries. They all face similar challenges around data, have opportunities that come with optimizing analytics and have to address the shortage of available data professionals.
In line with our key themes from Xperience 2019, I will review the panel discussion and share the key points with you here:
Data science in football
Data science is quickly gaining popularity in football, a development that is supported by the availability of large datasets for analysis and the affordability of computing power.
A lot of the research around certain performance metrics and the ongoing advancement of technology (including video and audio recordings of games), has given rise to innovation around AI in football and other sports.
With the data that’s now available and recorded at a rate of several dozen data points per player per second, clubs can ‘replay’ games purely based on the data – and they can analyze positions, angles, speeds, the length of a pass, and many other metrics.
At the same time, data science comes into play whenever data modeling is done. That can be in the area of recruitment, when a club is looking for a player with a certain profile and wants to use the available data from dozens or potentially hundreds of players to find the best fit – or a very promising outlier.
Modeling is also done to calculate the likelihood of certain events happening, which in turn informs things like player positioning and team formation.
The results of the work done by data scientists typically goes unnoticed by the spectators. It’s a long process which involves combining the insights derived from data analysis with the expertise and experience of the coaching team to find the sweet spot where an algorithm actually influences a coach’s decision on tactics. But it can make the difference between winning and losing.
The data resulting from a single game isn’t necessarily what you’d consider Big Data. But when you take multiple data sources and data types (e.g. tracking data plus events data) and you add multiple clubs, leagues and seasons, your data quickly grows in volume and you reach the limitations of Excel-based analyses.
Clubs, especially those from the big leagues in Europe and South America, are investing in building the right infrastructure to support these data volumes and to have the right tools for analysis – so they can gain the maximum value from their data assets.
With all that data, how do the data experts working for these clubs ensure that their stakeholders, (eg. the coaching team and other departments), get the right information at the right time?
As with businesses all over the world, the magic component is communication: coaches don’t need lots of raw data, they need information and insights that takes many factors into account and supports their decision-making.
Whether they need to understand injury patterns to inform their player selection for the next game, or address passing inaccuracies from a midfielder, it’s crucial for the data analysts and data scientists to have open communication channels for sharing the right information – and for understanding and anticipating what the coaching team needs.
Business intelligence and analytics
Data in football comes from various areas. On the one hand, there’s the player data, which tells the clubs about the performance of an individual player as well as the team, competitors.
But there’s so much more: data is also available for the marketing department, looking at ticket sales, merchandise and many other aspects of the fan and stadium experience.
And then there is the media with the data they use and need for their analyses, commentary and reports. Bookies wouldn’t be able to do their job and increase their odds without data to optimize their predictions.
Let’s stick with performance data, though, data from training and games. We saw previously that the data needs to be distilled into meaningful and actionable insights as well as communicated effectively.
That’s where analytics comes in. Behind the scenes there are people and whole teams focusing on this work, analyzing the data, putting it into context, connecting different data sets and finding reasons for certain events, as well as working on predicting future events. For example, those resulting from a certain pass or duel.
Clubs are using analytics tools to wrangle their data into meaningful reports, charts and dashboards that can be given to the coaching team.
Here, it’s important that these data visualizations are effective and visually impactful. Reports need to be easy to consume and understand, while showing very clearly what the key message is. While there is a range of options used by clubs, going from paper-based documents, to pdf files and fully-interactive dashboards presented on iPads during live games, the common priority is for simple and effective communication of information.
The main reason for using data visualizations to communicate information is that it shortens the time to insight and therefore enables faster decision-making and a shared understanding.
We often see data visualizations used in the media as well by ‘fanalysts’ ( a new generation of data-savvy fans) on social media as people explore publicly available data – and join the discussion about the game from a data perspective.
For instance, there’s a whole community that has shaped around analytics and data science in football, with high-profile bloggers and analysts commenting on the state of football regularly, and especially during big tournaments.
Cloud deployment options have become simpler and a genuine option for many businesses, including football clubs. Many clubs are embracing the flexibility of the cloud, but also take their responsibility for player privacy very seriously.
Data for medical records and other personal information remains in storage on-premises as clubs utilize hybrid approaches that give them flexibility while ensuring security and compliance with legal regulations.
In football, the player associations keep a careful eye on personal information such as biometric data, as part of their role of protecting the players they represent.
Summary – what next for data in football?
Data presents huge opportunities in football. Much of it is yet to be uncovered because football clubs, just like businesses everywhere, struggle with data challenges.
Clubs face disparate data sources and the need for better-integrated systems. While the data gathering, cleansing and preparation processes are time-consuming – taking focus and resources away from the value-added analysis work most data analysts and scientists love to do.
Then there is a gap to bridge between the ‘sports people’ (the coaching team, athletes and sports scientists) and the ‘data people’ (the data analysts and data scientists) to make sure that data, insights and information are understood, are actionable and are used effectively.
Communication is crucial in this process, as is looking beyond the boundaries of the sport and into different fields and industries for inspiration and ideas.
And finally, there is the question of what to measure and report. Certain things can be measured but aren’t meaningful or relevant, while some indicators that would be very insightful are extremely hard to measure with current technology.
What this panel discussion revealed to me is that football and the organizations involved in the sport face very similar data challenges to businesses and organizations in other industries. What it also showed me is that there is not just a lot of passion for the game and the data that exists behind the scenes, but also a lot of determination to overcome these data challenges. And there’s an openness and willingness to share ideas, innovations and successes with others to help everyone up their game.
In short, the beautiful game is football – and its future is data.