⚽️ Head of Data & Technology
@ClubBrugge
👨🔬 Research Fellow
@KU_Leuven
🎮 Data Editor for EAFC
@EASports
👨🎓 Computer Scientist and PhD in Machine Learning
With the growing interest in soccer analytics, it can be challenging to sift through all of the content that appeared in 2023. My latest blog post lists the research papers, blog posts, news articles, podcast episodes, and events that I liked the most!
🧵 Computing and visualizing advanced soccer analytics metrics is becoming extremely straightforward. Access to event data and basic programming skills suffice nowadays as free open-source software libraries take care of most of the heavy lifting for you! (1/20)
A ton of exciting soccer analytics content has been published in 2021. My latest blog post lists the research papers, blog posts, news articles, webinars and podcasts that I liked the most!
I am excited and proud to announce that I have joined reigning Belgian champions
@ClubBrugge
as Data Scientist. I am looking forward to helping drive the adoption of data throughout one of the world's most innovative football clubs.
With the growing interest in soccer analytics, it can be challenging to sift through all of the content that appeared in 2022. My latest blog post lists the research papers, blog posts, news articles, books, and podcast episodes that I liked the most!
The amount of soccer analytics content has spiked in 2020. My latest blog post lists the research papers, blog posts, news articles, events, invited talks, webinars, podcasts, Python libraries and newsletters that I liked the most!
Former Olympique de Marseille Data Scientist Benoit Pimpaud (
@Ben8t
) has published an interesting three-part series of blog posts on pursuing a career in football analytics. (1/5)
🧮 I computed the most likely opponent for each club in the UEFA Champions League round of 16. Taking into account that clubs from the same group and same association cannot face each other, my constraint solver produced 3,876 possible draws. ⬇️
#UCL
#UCLdraw
I'm currently hiring for a Match & Scouting Data Analyst role in my team
@ClubBrugge
. If you believe your skills and experience align with this exciting opportunity, then please apply through our website. ⬇️
The sports analytics event season is about to start! The Workshop on Machine Learning and Data Mining for Sports Analytics, StatsBomb Conference and Cascadia Symposium on Statistics in Sports are all taking place next week. Many interesting papers are available already! ⬇️ (1/4)
The "Football Analytics: Now and Beyond - A Deep Dive into the Current State of Advanced Data Analytics" guide that
@BarcaInnoHub
launched at the Barça Sports Analytics Summit last November is now available online.
Extremely excited about today's major update of the SciSports Platform! Our latest release includes a complete overhaul of the player profile, which focuses on providing additional context and presenting the take-away messages that emerge from interpreting our metrics. (1/5)
I'm collecting and aggregating data-driven predictions for the 2022 FIFA World Cup in the spreadsheet below. I'm primarily interested in model-based simulations. If you come across predictions that I could include, then please share them!
#FIFAWorldCup
Several exciting sports analytics conferences and workshops will be taking place later this year, some of which are still accepting paper and abstract submissions. These events are a great opportunity to get feedback on your recent sports analytics work! ⬇️ (1/7)
NEW | StatsBomb Customer Case Study
We spoke with Club Brugge's Football Data Scientist, Jan Van Haaren (
@JanVanHaaren
), about the impact of StatsBomb's data in the Brugge organisation and the club's title-winning season.
Read the full case study 🔽
Last month I had the pleasure of giving a keynote talk at
@BarcaInnoHub
's Barça Sports Analytics Summit. I presented a few event-data based football metrics that we developed for player recruitment. My presentation slides are available on Google Drive.
The limited explainability of expected value metrics for football is holding back their adoption by practitioners. Therefore, I am exploring ways to improve their explainability in a paper that I will be presenting at the AI for Sports Analytics workshop.
"SciSports, the global thought leader in data intelligence for the pro football industry, has partnered with Sports Interactive to integrate Expected Goals into the upcoming release of Football Manager 2021."
@SciSportsNL
@SI_games
@FootballManager
#FM21
Announcement: I resigned from my position
@SciSportsNL
. I am grateful for my time with the company and extremely proud of what we have achieved over the past four and a half years. I had the pleasure of working with many great colleagues and wish them all the best for the future!
"A high-level, data-driven sports analytics framework. floodlight is a Python package for streamlined analysis of sports data. It is designed with a clear focus on scientific computing and built upon popular libraries such as numpy or pandas."
Since last season, Hawk-Eye has been providing
@ChampionsLeague
participants with skeletal tracking data. Last season clubs had to pay an additional fee to obtain access to the data, whereas this season the data is available free of charge. (1/3)
Our 8th Workshop on Machine Learning and Data Mining for Sports Analytics took place last Monday. The program included a great invited talk on "Valuing Actions and Assessing Players" by
@schulteoliver1
and 17 contributed talks. The papers are available on .
"The Eredivisie is making a fresh step in collecting and sharing sporting data relating to Eredivisie matches. Tracking and event data will be supplemented with special video footage and all data will be brought together on a central analysis platform."
Several exciting sports analytics conferences and workshops will be taking place this summer, some of which are still accepting paper and abstract submissions. These events are a great opportunity to get feedback on your recent sports analytics work! ⬇️ (1/5)
As a result, I am currently looking for a new challenge, preferably within football analytics or sports analytics. If you are interested in having a chat with me, then please feel free to get in touch!
If you're interested in learning more about football analytics but find research papers daunting, then I can highly recommend Alex Marin Felices' blog (
@alex11marin11
). He condenses influential research papers into easily digestible summaries.
I have started collecting data-driven
#EURO2020
predictions in this spreadsheet. I am sure that I am still missing many predictions. If you know of any predictions that I could include, then please drop me a message!
I'm proud to announce that our paper "Actions Speak Louder Than Goals: Valuing Player Actions in Soccer" has received the Best Paper Award for the Applied Data Science Track at the 2019 ACM SIGKDD Conference on Knowledge Discovery and Data Mining!
#KDD2019
#KDD19
Last week at the
@StatsBomb
Conference, I was glad to learn that my annual review blog posts had been used by the authors of several research papers that were presented on the research stage. As has become tradition by now, my next review blog post will appear in late December!
The artificial intelligence hype in the soccer industry is clearly getting out of hand. If the trend continues, we should soon have more start-ups that claim to be able to provide "insights from data using artificial intelligence" than actual professional soccer clubs.
If you are new to football analytics, then you should definitely look into kloppy. This Python library is extremely powerful but unfortunately does not receive the attention that it deserves.
Super proud to announce kloppy version 2.0.0 🥳
This (breaking) change includes:
1. Default normalization of coordinates (read more )
2. Add missing statsbomb shot outcome values
Big thanks to
@brunodagnino
and
@mbordese
!
The UEFA Champions League group stage draw takes place tonight. Since I was wondering how the draw regulations affect the chances of clubs ending up in the same group, I wrote a Python script that simulates the group stage draw procedure. My code is available on GitHub. ⬇️ (1/4)
I'm proud and excited to be appointed Chief Product & Technology Officer
@SciSportsNL
! I'm very much looking forward to help shape the future of our football analytics tools using state-of-the-art machine learning and software engineering techniques together with my great teams!
The
@scikit_learn
core developers are running another edition of their free course on machine learning in Python. The course discusses many concepts that are relevant for sports analytics. (1/2)
The 9 research paper finalists and 11 posters for the research-paper competition at the 2020 MIT Sloan Sports Analytics Conference have been announced.
#SSAC20
#SSAC
@SloanSportsConf
The
@PySport
open-source website () provides an excellent starting point. The website currently lists no fewer than 2 open-data repositories and 44 software libraries: 24 Python libraries, 19 R libraries and 1 Haskell library. (2/20)
The seven papers that have made the finals of the research-paper competition at the MIT Sloan Sports Analytics Conference are now available for download from the conference website.
#SSAC21
Proud that my joint research with
@LotteBransen
on quantifying player chemistry in football has emerged as runners-up in the prestigious research-paper competition
@SloanSportsConf
! The competition received over 200 papers across all four tracks.
#SSAC20
I just finished reading Football Hackers by
@chbiermann
following a suggestion from my colleague
@rob_seidl
who had read the original German edition of the book. If you're into football analytics, then this book is definitely a must-read! (1/5)
Tottenham Hotspur's (
@SpursOfficial
) potential opponents in the UEFA Champions League round of 16. ⬇️
#UCL
#UCLdraw
1. Inter Milan - 18.9%
2. Borussia Dortmund - 18.0%
3. RB Leipzig - 17.4%
4. Milan - 17.3%
5. Club Brugge & Paris Saint-Germain - 14.2%
Yesterday
@TomDecroos
successfully defended his PhD after four years of hard work! I had the honor and pleasure to serve as the secretary of the examination committee. In his excellent presentation, Tom highlighted three key contributions to the field of soccer analytics. (1/5)
Four years of my research condensed into this little book... For those who want a 194-page deep dive into soccer analytics (the intro and background section could be interesting for people just getting into the field), here is the full text:
Excited to announce that the Contribution Ratings module in the SciSports Platform has been extended with metrics that assess the impact of defensive actions! The following blog post provides further information on these cool new metrics. (1/4)
Our paper on optimizing on-pitch decision-making in soccer has been selected for the final of the research-paper competition at the MIT Sloan Sports Analytics Conference.
#SeizeTheData
#SSAC22
Today I had the pleasure of speaking at Training Ground Guru's 2021 Big Data webinar. Many thanks to
@ground_guru
and
@HugoScheckter
for having me and hosting this wonderful event!
#BigData2021
When
@johnspacemuller
and I talked about the use of data and analytics in player recruitment back in early May, I was still in the dark about my future. I could never have imagined that today I would be implementing some of the ideas that we talked about inside a wonderful club.
The past month has been surreal. I have had plenty of interesting conversations with clubs, companies and academics in the sports analytics space. I am extremely grateful for the massive support that I have received. Thank you to everyone who helped spread the word!
Announcement: I resigned from my position
@SciSportsNL
. I am grateful for my time with the company and extremely proud of what we have achieved over the past four and a half years. I had the pleasure of working with many great colleagues and wish them all the best for the future!
As
@Soccermatics
announced last night,
@LotteBransen
and I will discuss our work on valuing on-the-ball actions in Friends of Tracking on Thursday May 7th. We will also record a few videos. If there's anything in particular that you'd like us to discuss, please let us know!
#FoT
"We created action embeddings — a representation of football actions; and player embeddings — a representation of on-the-ball actions of football players. (...) We can use the model to find players with a specific style of play."
I have started preparing my annual review blog post that summarizes my favorite soccer analytics content from the past year. If you have produced or come across content that you believe deserves a spot, then please share the link via my Linktree page!
Our AI for Sports Analytics (AISA) workshop at IJCAI 2021 is open to sports analytics papers that involve machine learning, game theory or computer vision. The full call for papers is on the workshop website and the submission deadline is May 5th.
I'm happy to announce that the sixth edition of our Workshop on Machine Learning and Data Mining for Sports Analytics will take place as part of the European Conference on Machine Learning in September. Further details are on the workshop website: .
In their latest publication,
@Florisgoes
et al. analyze the tactical characteristics of successful attacks using tracking data for 118 Dutch Eredivisie matches. Their research is now available as open access paper.
Today
@p_robberechts
will present his research on contextualized performance projections for football players at the 2020 OptaPro Analytics Forum. If you are at the event, then make sure to check out his excellent poster!
#OptaProForum
@OptaPro
I'm thrilled to announce that last week, I accepted an offer to join Club Brugge as a Match & Scouting Data Analyst. This marks a significant milestone in my career, and I'm deeply grateful for the opportunity to be part of the most prestigious club in Belgium.
The video recording of
@LotteBransen
's great talk on quantifying player chemistry in football at the 2020 MIT Sloan Sports Analytics Conference has appeared online!
#SSAC20
@SloanSportsConf
(1/2)
Excited about the Reinforcement Learning competition that
@GoogleAI
and
@ManCity
are hosting on Kaggle!
"In this competition, you’ll create AI agents that can play football. Teams compete in steps, where agents react to a game state." (1/4)
Our 9th Workshop on Machine Learning and Data Mining for Sports Analytics (MLSA) at ECML/PKDD 2022 is open to sports analytics papers that involve machine learning and data mining. The submission deadline is Monday June 20th.
@ECMLPKDD
(1/2)
On this day in 2016, I successfully defended my PhD dissertation titled Relational Approaches for Learning, Transferring and Mining. Browsing through the slide deck that I used back then made me realize that I should give relational machine learning techniques another shot! (1/2)
While large language models continue to grab headlines, there's been little news on the soccer front. I'm intrigued by what other sports clubs are up to behind the scenes, and whether we can expect any commercial products that effectively leverage this technology in the future.
Honoured to announce that
@Soccermatics
will give an invited talk titled "Using collective motion models to help players improve their game" at the 7th Workshop on Machine Learning and Data Mining for Sports Analytics on Friday!
@ECMLPKDD
#ECMLPKDD2020
(1/2)
The recording of my talk on "Evaluating & Recruiting Players Using Open-Source Tools" at Training Ground Guru's Big Data 2021 is now available via the
@ground_guru
website.
Today I had the pleasure of speaking at Training Ground Guru's 2021 Big Data webinar. Many thanks to
@ground_guru
and
@HugoScheckter
for having me and hosting this wonderful event!
#BigData2021
For the past few years, I have maintained a list of soccer analytics Twitter handles that share relevant content. If you believe that a particular handle is missing or should no longer be in the list, please let me know. I am open to suggestions.
The 7th Workshop on Machine Learning and Data Mining for Sports Analytics takes place today. The program features an exciting invited talk by
@Soccermatics
in addition to 12 contributed talks.
@ECMLPKDD
#ECMLPKDD2020
(1/2)
We will present our paper "Actions Speak Louder Than Goals: Valuing Player Actions in Soccer" at KDD 2019 on August 7th. If you cannot attend our talk, you might want to check out our video. We need your views to win the video competition. Please RT! 😉
Congratulations to
@EightyFivePoint
and
@suds_g
for winning the research-paper competition at the 2021 MIT Sloan Sports Analytics Conference with their paper titled "Routine Inspection: A Playbook for Corner Kicks"!
#SSAC21
@SloanSportsConf
I've been running a sports analytics reading club at our Amersfoort office for the past year. We meet every three to four weeks to discuss a relevant paper. I'm considering to occasionally open up our reading club to external people. Would anyone of you be interested in joining?
I had a great day at the Barça Sports Analytics Summit! Many thanks to
@JaviOnData
and
@BarcaInnoHub
for inviting me! I will still be around for the Barça Sports Tech Symposium. If you would like to talk about analytics, then feel free to drop by!
#BarcaAnalytics
#SportsTechWeek
Check out a novel
#ReinforcementLearning
environment where agents aim to master the world’s most popular sport—football! The Google Research Football Environment includes benchmarks & progressive RL training scenarios, and is available in open source beta→
Congratulations to Gabriel Anzer,
@pascal_bauer
, Ulf Brefeld and Dennis Fassmeyer on winning the research-paper competition at the MIT Sloan Sports Analytics Conference!
#SeizeTheData
#SSAC22
Congrats to Detection of Tactical Patterns Using Semi-Supervised Graph Neural Networks for placing 1st in our Research Papers Competition 🏆
#SSAC22
#SeizeTheData
I'm on my way to London for tomorrow's
@StatsBomb
conference together with
@jessejdavis1
. If you're interested in having a chat, then don't hesitate to get in touch at the event or via Direct Message!
The UEFA Champions League group stage draw takes place tonight. Last season, I wrote a Python script that simulates the draw procedure. I now updated the script with the clubs, pot assignments and club pairings for this season. My code is available on GitHub. ⬇️
This thread is full of great advice for aspiring data scientists. Like software engineering a few decades ago, data science is now a fast growing and young field.
"If you're on the younger side, also don't cheat yourself. Find you a team with experienced leaders to learn from."
Chrome was delivered without any sprints at all. The team came in at 9 and left at 5 (figuratively, people actually kept their own ~8h schedules) every workday for a couple years like clockwork. No drama. No broken marriages, no broken families.
The SoccerAction library is the main work horse. I use this library to convert event data from different providers into the SPADL and Atomic SPADL formats, which are unified representations of on-the-ball player actions that facilitate analysis. (4/20)
Excellent tutorial on learning from positive and unlabeled data by
@jessa_bekker
. This learning setting is highly relevant for many sports analytics tasks.
Our 8th Workshop on Machine Learning and Data Mining for Sports Analytics (MLSA) at ECML/PKDD 2021 is open to sports analytics papers that involve machine learning and data mining. The submission deadline is June 23rd.
@ECMLPKDD
Workshop website: .
I just received my copy of The Ten Equations that Rule the World by
@Soccermatics
. My expectations are high as I really enjoyed reading his earlier books Soccermatics and Outnumbered!
The 11th Workshop on Machine Learning and Data Mining for Sports Analytics (MLSA) will take place in September in Vilnius, Lithuania. Our workshop welcomes sports analytics research papers involving machine learning and data mining.
@ECMLPKDD
(1/2)
If you have ten minutes to spare, I recommend reading
@EveryTeam_Mark
's latest newsletter on how football clubs spend their analytics money. You will not regret your choice.
I've migrated my personal website from Jekyll (
@jekyllrb
) to Quarto (
@quarto_pub
), including my annual soccer analytics review blog posts. In case you've bookmarked any of the blog posts, your bookmarks should still work as I've set up redirects from the old to the new locations.
Last night
@p_robberechts
talked about valuing on-the-ball actions (VAEP) and quantifying the impact of mental pressure on player performance at a
@dsleuven
meetup on sports analytics. A recording of his excellent talk is available on YouTube.
I'm collecting and aggregating data-driven predictions for the 2023 FIFA World Cup in the spreadsheet below. I'm primarily interested in model-based simulations. If you come across predictions that I could include, then please share them!
#FIFAWWC
In this blog post,
@laudevs
,
@p_robberechts
,
@wannesm
and
@jessejdavis1
show that action value models based on boosted trees are susceptible to adversarial examples, where slightly perturbing the feature values leads to vastly different action values.
I have recently joined the
@kloppy_dev
project as a co-maintainer alongside
@mr_le_fox
! This open-source Python library aims to facilitate the process of loading and processing soccer data. I am hoping to contribute to this exciting project in the following three ways. ⬇️ (1/5)