Visit Homepage
Skip to content

Shake those tail feathers (3)

All good things come to an end, and so it is with the 20th edition of the World Cup. After a month full of love bites, nerve-wrecking extra time, Schwalbe’s that could teach this kid a lesson or two, and even a game where you only needed to blink to miss a goal, we can finally catch up on that much-needed sleep and return to a more healthy dietary regime than potato chips or ready-to-eat meals (running back and forth between the TV and the microwave during those 5 bloody minutes, ready-to-eat my ***! I practiced so much on that incendio spell with my wand; I guess I should really start accepting the fact that Muggle material’s all I’ve got in me, and that Hogwarts doesn’t send out acceptance letters to 28-year-olds…).

So, this is what the final network looks like:

Germany clearly stands out here, with the highest weighted outdegree. According to our statistics, they should have faced the Netherlands in the final instead of Argentina (based on the number of goals in the previous games), but those ****** Argentinians with their sterile defence again (we’re not a fan here in Belgium…). I wasn’t quite satisfied with this graph though. With the draws left out, it kind of splits the tournament up into two divisions and with that giant arrow pointing at Brasil somehow creates the illusion that that game was the grand finale (though in terms of jaw-dropping statistics it probably was a winner).

So I decided to try the old method again and incorporated all matches, with the draws as undirected ties. This is the result:

Since not all edges are directed, Gephi won’t compute the outdegree so I couldn’t size the nodes accordingly. But the group stage comes out more clearly here, and the final and play-off for third place are nicely grouped in the center here.
At the NY Times they looked at the teams from another angle, by checking out which clubs are best represented in the national teams. Seems to be some kind of three-mode network (country – player – club), never seen a graph like this before. Yet another example of the cool stuff you can do with network analysis! Or not, if you’re not into soccer…
Recap of the historical sessions at EUSN coming soon!
Love you long time,
Your Data Ninjas

UPDATE: more World Cup statistics! Facebook’s data science team analyzed shifting fan support during the different stages of the tournament. Check it out here!

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *