_images/SoccermaticsLogo.png

Welcome to Soccermatics

This course is the most comprehensive education available on how to work with football data. We will learn how to understand the game using mathematics, statistics and machine learning.

It is taught by Soccermatics author David Sumpter, with guest lectures from world-leading practitioners and researchers working with data in football. It offers the technical skills you need to get a job as a data scientist at a football club or in the football industry. Students graduating this course have found jobs in all parts of the football industry. And clubs and national teams send their data scientists to this course in order to improve their skills.

The course utilises data from Wyscout, Statsbomb, Metrica, Signality, Skillcorner and other leading data providers. It also reflects the state of the art in research on football analytics. Much of the work presented is based on recent work presented at, for example, Sloan Sports Analytics and The Opta Forum.

It is also a fun and interesting way to learn mathematical modelling techniques used in many areas. Indeed, this course isn’t the same as a normal maths, statistics or machine learning course, where you might learn specific techniques. The type of techniques we will use vary alot. We will start with techniques from data visualisation, then move on to logistic regression, learn about various probability distributions, a bit of geometry and physics, self-propelled particles and use some machine learning and AI methods.

The reason for this structure is that the real world is not like a maths class. When we model reality – be it a football match, the spread of disease or the movement of a bird flock – we don’t start the process on page one of a textbook. Mathematical modelling, which this course is about, is more creative than that. For each problem we face, we need to use a different piece of mathematics to get to the heart of the problem.

Sometimes very advanced topics in maths and statistics, turn out to be much easier to grasp in the context of applications. Probability theory can be quite an abstract area of maths, but by looking, as we do in section 4, at a concrete example of goals in football or soldiers getting kicked by horses, we can quickly get a feeling for the theory. Abstract ideas become more concrete with applications.

Three ways to do this course

Students: This is 5 credit course at Uppsala University and is taught as an online course in late-August to end of October each year. It is free for all EU citizens and available as a paid course for those outside the EU. The university includes tutorials with me and a teaching assistant, individual hand-ins and group work. Application to participate in the course opens in March/April each year (i.e. half a year before the course start). Late registrations are sometimes available.

Professionals: If you are a part of a footballing organisation, then it is possible to take this course as part of further education. The professional course has less formal requirements for entry, allowing your organisitaion to benefit from it. We help participants to work with their own data on the assignments. The course ran for professionals ran twice in 2023 and was attended by data scientists from Europes big leagues (including sevral from the Premier League), national organisations and clubs from all over the world.

It will run again in September 2024 and 2025. The cost is €1600 per person (discounts if several participants from same organisation). It is run by Twelve Football where you can find more info and register.

Non-enrolled students: Several students (some of whom are now working for football clubs) have completed this course through online study groups. To do this, you need to find a group, since all final projects are in groups. If you would like to do this, please go in to the Slack channel and find some like-minded people to work with.

If you complete all three assignments for the course message David Sumpter in the Slack channel, send him your assignments, and if they are complete then he can send you a certificate saying you have completed the course. Unfortunately, I am not able to provide feedback for non-enrolled students along the way. But often by working together in the Slack group, dedicated students find a way of learning together.

If you pass the Uppsala course, the professional course, or are issued with a certificate from the non-enrolled course, it is because you have shown the skills needed to be a football data scientist. This means that clubs can rely on the certificate when hiring data scientists.

Required Background

When I teach this course at Uppsala University, we set prerequisits of university level courses in statistics, linear algebra and calculus. It is taught as a late undergraduate/Masters level course. There will be starter/refresher Intro to Python programming and statistics, but the learning curve will be relatively steep for those without this background. In short, to do this course you should have studied mathematics and programming at University level. You should also, of course, have an interest in football.

For the professional version of the course you should work as data scientist, match analyst or data analyst, with the ambition and drive to learn to analyse football with data.

It is possible to grasp the concepts without the technical background, especially if you love football! So, feel free to give it a try. The worst that can happen is that you will learn to plot shot locations on a football pitch.

History

This course developed from the Friends of Tracking Youtube channel. It was recorded during the Covid-19 lockdown in April/May 2020 by a group of club analysts and leading academics. These include lead data scientists: Javier Fernandez (Barcelona), Sudarshan Gopaladesikan (formerly Benfica, now Atalanta) William Spearman (Liverpool), Laurie Shaw (Manchester City), Fran Peralta (formerly Hammarby, now Atletic Bilbao), Alex Thomas (The English FA), Pascal Bauer (German DFB) and many more. We initially made this work available at the Friends of Tracking Youtube channel.

This course uses some of those videos and adds further videos and course notes. It was made possible, through the kind help of Andy Rowlinson, who both allowed us to use MPL Soccer (a library for plotting event data) and helped set-up this website. The code for this course was written by David Sumpter, together with Aleksander Andrzejewski (a former student of the course). We use several examples in this course from work with twelve.football .

_images/History.png

Soccermatics

This course also builds on and uses a number of examples from my book Soccermatics . We will often return to the book for inspiration, like the opening the paragraph of the book below,

_images/SoccermaticsIntro.png

But we will also go further than the book. Indeed, the reason I have created these pages is to teach the maths behind Soccermatics. To give you the analytical skills you need to understand the beautiful game and to create mathematical models. This course is (I think) more fun if you work on it while you read the book. So don’t delay, get it today.

Since writing Soccermatics, I have worked with data scientists at the English FA, Barcelona, Manchester City, Benfica, Ajax and Hammarby. I have worked with sporting directors, scouts, managers, TV presenters, and even players with data analysis. I will bring many of those experiences to the course.

_images/Me.png

Buy one of David’s books

The best way to support David’s time devoted to creating this course is to buy one of his books, either for yourself or for a friend or family member.

_images/MyBooks.png

These are available from Your Local Bookshop or from Amazon

Buy the T-shirt!

If you are studying the course and already have the books, you might as well buy the t-shirt! Or get a mug, a water bottle or another Soccermatics themed item.

_images/BuyMerch.png

These are exclusively available from Beautiful Equations

Contents