Being the best team in the world is not easy. It means that every team who faces you will run out to the pitch with an extra bit of motivation, with extra hunger to perform and beat the greatest. But it’s also a responsibility. Being the best means coaches and analysts will be all around you. They will analyze the smallest details to learn your style, model it, and then try to recreate it for their own team. But which teams are already playing like the champions of Europe? The visualization below tries to answer this question.
To create this tool, I used 25 different metrics from Wyscout, and put them into a K-Means clustering model. K-Means clustering is an unsupervised learning algorithm that, as the name hints, finds a fixed number (k) of clusters in a set of data (A cluster is a group of data points that are grouped together due to similarities in their features). After creating the clusters I measured the correlation-based distance between all the 675 teams in the dataset. Lower distance indicates a more similar team, while a higher number tells us that there is not much in common between the measured teams playing style.