Open catterbu opened 9 years ago
@catterbu, this needs to be rebased off of my dev before I can merge, in effect removing commit 2eeefb0 (Determine k_best using NbClust
). I can't merge that commit yet, it should only be in PR #28.
Sorry, not in effect removing that commit, what am I saying. You have to remove the commit by hand from this branch, rebasing won't do the trick.
Nice job on this. I wouldn't focus on this comment now, but in 3D it gets hard to interpret a static image with more than a few points:
I think we might want to consider holding further development on any static 3D visualizations and aim to do all 3D stuff in JS. If you can get the coordinates of each of the points in the PCA we can build something like this very quickly: https://jsfiddle.net/rd26u1zj/1/. Let me know if possible.
@nsh87 I just pushed what should be the final version of the pca plot stuff. All the rebasing and everything should finally lead to seemless from this point. the boxplot is ready too, but I will need to rebase off dev after this is merged before that one can be too.
What changes are in the box plot?
it will be a separate pull request shortly. that is the one for handling boolean data.
I don't think you'll need to rebase that off dev if I merge all your stuff first/in the order of your commits.
This adds a 3D PCA plot called
pca_plot3d
with a wrapper onscatterplot3d
. Also, the title for both PCA plots refer to the amount of variance explained.