Open jessexknight opened 2 years ago
I would absolutely welcome assistance with this. Please feel free to experiment and submit PRs.
The trouble is not with performing the statistical transformations or parameterizing the graphical objects but with playing nicely with the interdependencies of ggplot2's builder. I would love to see a wiring diagram (like this) of the process, which would help create a workable solution. Though it's been a while since i looked into it deeply, and i can't say from memory specifically where group
prevents a problem.
(Though it looks like gganimate also co-opts the group
aesthetic, presumably because it faced a similar problem. If a workaround could be found there, it would resolve the conflicts both with ggalluvial and with ggforce.)
Possibly related to this issue, I think maybe hijacking the
group
aesthetic internally result in incompatibility with the gganimate package. Maybe some other people are also affected. Not sure how much work / bugs it may introduce to change this use ofgroup
internally, or whether you could suggest a workaround?