178 players were present for at least one point scoring moment. Wales used the most players with 34, England used the least with 27 players (you can't say Eddie Jones doesn't know his team). I was expecting France to be the team that used the most players since France were the team with a COVID outbreak.
The point-scoring moments were reasonably well spread out in time, with interesting peaks at 18 and 65 minutes.
No minute had a point scored by all team, but several had points scored by 4/6 teams - including the 18th but not the 56th minute.
Comparing who scored the points for the 6 teams is interesting:
All the teams are dependent on kickers, as expected. It's possible that Scotland's is less skewed towards Finn Russell than it might be otherwise because of the game time Stuart Hogg got, due to injury or Russell being sent off.
Ireland are that badly skewed towards Sexton even though Sexton missed a match. This does not bode well for when he retires.
Players who were on the pitch at the same time:
This is the players who were on the pitch for more than 28 point scoring moments each. Yes, Rees-Zammit and Tipuric stand out even in this field.
I tried to make a dendrogram of all the teams. To make the labels readable, you need to reduce it to only those players on the pitch for at least 25 point-scoring moments.
Unfortunately this removes all the Italian players.
To include any Italian players, you have to drop the cut-off to at least 13 point-scoring moments.
The same is true for the heat maps.
Below is one where the labels are legible (all players were present for at least 25 point-scoring moments)
Below is the one that includes any Italian players.
Comparing the heatmaps for the different teams is interesting:
They're such different patterns. I mean, there's some similarity between all of them but, for instance, the Scottish one is much less dense, while the colouring for "on the pitch at the same time" is tightly packed in one corner for Wales. For Ireland it's packed, but less tightly. England is packed, but more in the middle-ish. I'm reasonably sure I've seen a carpet with the same pattern as France's and Italy's would make a good start for an artwork.
Conclusion: This has definitely been worth doing, not just from a "learning how bits of R and json work" point of view. Visualising the data has revealed some interesting patterns, which I think reflects things about the teams. It would be interesting to see how things like this look in sports with rolling subs (odd that there's a Rugby League World Cup this year that might fill that gap ;) )
Other possible future work could include doing the same thing but looking at which players are on the pitch when points are conceded (although that could be unfair to Italy), or looking at attacking play again, but in the 2022 Six Nations, because there's only been one coaching change in the interim (Italy) and it'll be interesting to see how many of the patterns repeat - Ireland's reliance on Sexton, the very centralised England core, France not scoring (many) late tries. This analysis may be affected by England's enforced change of kicker due to Owen Farrell's injury.
