kde plot `'y'` argument ends up being 'x' label - rename it to `columns`?

[MarcoGorelli]

ALL software version info

pandas: 2.1.4 hvplot: 0.9.1

Description of expected behavior and the observed behavior

The 'y' argument determines which column(s) to compute distributions of.

It's a bit counter-intuitive, then, that the x-label ends up being the y argument

Don't know what the simplest solution is - rename y for this plot, perhaps to columns?

Complete, minimal, self-contained example code that reproduces the issue

df = pd.DataFrame({'a': [2,4,2,5,6,7,3,6,7]})
df.hvplot.kde(y='a')

Stack traceback and/or browser JavaScript console output

Screenshots or screencasts of the bug in action

• [x] I may be interested in making a pull request to address this

[MarcSkovMadsen]

+1 :-)

And if for some reason renaming is not accepted. Please document why y makes sense to use.

[MarcoGorelli]

Please document why y makes sense to use.

I think there is a valid interpretation that here you're plotting a kernel density plot, and the y-axis ends up holding the density of column y

Can look confusing though

[maximlt]

It turns out this is Pandas' API:

The difference is that hvPlot shows a in the x-axis while Pandas doesn't.

I'm okay with how hvPlot displays the plot by default. Seaborn displays it the same way:

You'll notice they use x instead of y 🙃

Please document why y makes sense to use.

So I would say hvPlot should better document why y should be used, not to divert from Pandas' API.