Women's World Cup 2023 - Final Network Diagram

As you can all imagine, full working decorum was maintained at all time between the Australian and English offices for the entirety of Wednesday morning (UK time) and the company messaging system was full of GIFs flying in all directions. 

(We also have the rugby union and cricket World Cups where similar battlelines might be drawn.) 

The network diagrams now look like this: The club team closest to the centre are Manchester United. 

Because so many of Spain's players play in Real Madrid and Barcelona, but England have two players playing for Barcelona, England are slightly closer to the centre. 

Barcelona are the team with the most representatives in the final, with 10 players. Next are Real Madrid with 8 and Manchester City with 6. 

Manchester United and Barcelona link the two national teams, with Ona Batlle of Spain, and Mary Earps, Ella Toone, Katie Zelem and Alessia Russo of England playing for Manchester United and Irene Paredes, Aitana Bonmati, Mariona Caldenty, Alexia Putellas, Laia Codina, Maria Perez, Salma Paralluelo and Cata Coll of Spain and Lucy Bronze and Keira Walsh of England playing for Barcelona.

Once a tournament is down to two teams, the community view doesn't give a lot of information, but I'm including them here because it's quite pretty this time.  

I will be spending Sunday morning hiding from the match so I don't jinx England. I have been informed by colleagues that this is ridiculous. They will thank me if England win.

Women's World Cup 2023 - Semifinals Network Diagrams

In the predictions made in the last post, the diagram got 2/3 correct where there was a proper prediction.

Following the eliminations after the quarterfinals, the network diagrams now look like this:  

It's now sort of squished diamond shape, with most of the weight in the Australia, England and Sweden cluster at the bottom of the diamond. Spain are the team sticking out at the top. Sticking out is normally a bad sign for the next match, but Spain are Spain and are somehow doing well despite a whole lot of problems. 

England are the national team closest to the centre, with Chelsea (just about) being the club team closest to the centre. 

The club teams with the most representatives are Barcelona with 11 players left representing them, followed by Manchester City with 10 and then Chelsea and Real Madrid with 8. France going out wiped out the Paris Saint Germain and Lyon players (not totally, but not far short since it reduced them to 1 player left each). 

The community view is interesting, given there's 4 teams left and 5 communities. Arsenal are the mysterious 5th community, I think because they link 3 of the teams (Australia, Sweden and England). 

As for the diagram's predictions for the semifinals, they are as follows: 

Sweden vs Spain - diagram probably says Sweden, football knowledge says Spain. The diagram has been mysteriously right about Sweden so far but I don't think it can continue. 

England vs Australia - knowledge (esp. since Australia are at home) says "dunno", diagram says England just.

Women's World Cup 2023 - Quarterfinal Network Visualisations

I start with an admission: of the predictions I made in my last post except for the USA vs Sweden match, where the diagram disagreed with footballing knowledge, the diagram was wrong. 

But I'm still going to carry on making the network diagrams because I enjoy it and because they look pretty. 

The community views look particularly pretty this time, but let's start with the usual network diagrams that look at the links between teams.   

The club team closest to the centre is Manchester United, while France and England are the national teams closest to the centre. 

Denmark and Norway being eliminated has broken up the Nordic+Australia pack at the bottom. There's now a cluster of 5 teams (Netherlands, France, England, Sweden and Australia) with Spain and Colombia sticking up and Japan sticking out at the bottom. 

The club teams with the most representatives are Manchester City (with 12 players representing them), then Real Madrid and Barcelona (with 11) and Chelsea and Paris Saint-Germain (with 9). 

As I mentioned, the community views are very pretty.  

There are 8 teams left, who are now each their own community. 

For the quarterfinals, the diagram's predictions are as follows: 

Spain vs Netherland - Knowledge and diagram says "ooooooh" because that should be a good match. 

Japan vs Sweden - Diagram says Sweden, knowledge says Japan, Japan's players tending not to play overseas, is a thing, and that will affect the diagram. 

Australia vs France - Diagram says France, Knowledge says France are mis-firing but have depths of talent and are mercurial France! vs home town Australia so, who knows. 

England vs Colombia - Diagram and knowledge say England, although they've been making tough work of it so far.

Women's World Cup 2023 - Last 16 Network Diagrams

The first set of diagrams (https://fulltimesportsfan.wordpress.com/2023/07/22/womens-world-cup-2023-group-stage-network-diagrams/) was remarkably good at predicting the outcome of the group stages. 

The 3/16 that I didn't predict were Canada, although I did say group A would be tight, Italy, because they confounded me (again) and Germany, but I don't think anyone guessed that Germany would collapse like that. Definitely not German colleague who started a meeting with "please no jokes". 

To an extent, I think both Italy and Germany's closeness to the tight group at the bottom of the previous diagram was what meant I thought they'd go through when reading the diagram. 

With half the teams eliminated, the network diagram now looks like this. Much like Eurovision, there's a European and Australian cluster, which means the "central weight" of the diagram is not in the middle. 

That means that saying that either Leicester or the Chicago Red Stars are the club team closest to the centre, and either England or Nigeria are the national team closest to the centre doesn't give as much information as usual. 

The club teams with the most representatives left are Barcelona (with 15), Manchester City (with 13) and Paris Saint-Germain and Real Madrid (with 12 each). Manchester City and Real have crept up by not losing players, while other teams did. 

The community view looks like this: One day I'll figure out how to stop Gephi from cutting off the edges of the labels when the exceed the edges of the diagram. 

There are still three multi-team clusters, they are: 

England and Australia 

Denmark and Norway 

France and Netherlands 

These were also clusters in the first set of diagrams, with the change being that Sweden has moved out of the England and Australia cluster. 

While some of the eliminations were relatively straightforward to predict, the last 16 provides a number of games where "football knowledge" suggests one winner, and the diagram another. They are: Switzerland vs Spain - football knowledge says Spain, the diagram is unclear 

Netherlands vs South Africa - knowledge and diagram say Netherlands 

Japan vs Norway - knowledge says Japan, diagram says Norway 

Sweden vs United States - knowledge says United States, diagram says Sweden (US have been an outlying team, even when they win [https://fulltimesportsfan.wordpress.com/2019/07/01/womens-world-cup-semifinals/] so I think this might be one of the times when their is a weakness of the diagram based on circumstance. 

Australia vs Denmark - Ooooh, that's down as either way whichever method you look at it. 

France vs Morocco - knowledge and diagram say France, goalie hopefully says no (although I am happy with either team winning because I <3 renard="" wendi="">England vs Nigeria - knowledge and diagram says England Colombia vs Jamaica - knowledge says Colombia, diagram says Jamaica 

So this should be a good test of the underlying theory. 

(Even if it's wrong, I'll still keep making the diagrams because they're so pretty)