search

so-wise / weddell_gyre_clusters

Unsupervised classification of Weddell Gyre profiles
MIT License
2 stars 1 forks source link

Consider replacing the 3D PCA plot with a set of 2D plots #85

Closed DaniJonesOcean closed 2 years ago

DaniJonesOcean commented 2 years ago

This would make it easier to plot the Gaussian ellipses on top of the data. We could also use a "histogram" type plot too, which could be easier to see.

The 3D PCA plot is a projection anyway. So why not choose a projection that's easier to read?

DaniJonesOcean commented 2 years ago

https://matplotlib.org/3.5.0/gallery/statistics/confidence_ellipse.html

DaniJonesOcean commented 2 years ago

https://github.com/so-wise/so-fronts/blob/main/src/plot_utils/ellipses.py

DaniJonesOcean commented 2 years ago

This is really not a priority for the paper. Let's get it out there and worry about this another time.