In #6338 we discussed whether we want plotting code in sklearn.
We decided that the answer is yes, and we add a plotting submodule.
I think it would be good to discuss what functions to add.

Here are some that come to mind:

• better 2d scatterplot (with legend) - I have that for my book
• confusion matrix (have that too)
• plotting coefficients from a linear model or feature importances (maybe? It's relatively simple)
• plotting results from grid-search
• plotting learning curves
• plotting 2d decision functions
• plotting trees in the notebook (currently getting them into a notebook is a bit of a pain, though the notebook actually can plot trees directly)

Not sure if there are other common types of plots. Any input welcome.

Also: we have quantile plots that we might want to move?