Improve the API of the hvPlotExplorer

There are mostly suggestions that occurred to me while writing the release blog post:

• by providing an easy entry point such as hvplot.explore(df) instead of having to instantiate the class hvPlotExplorer. I wasn’t convinced by the name hvPlotExplorer either but it’s fine if there’s an easy and short entry point instead.
• by providing a way to get a handle on the plot. Should maybe get inspiration from Interactive that defines methods to get a handle of the various outputs it offers, such as .widgets(), .panel(), .holoviews(), etc… Here I’d say I expect a .plot() method at least.

[maximlt]

Yes it’s interesting because I think when I read hvPlotExplorer I understand that it is to explore data, while in practice I would more likely use that component to edit a plot. In that case I already have an idea of the kind of plot I want to get, it’s just that the .hvplot API has so many parameters that building a plot from this component makes my life easier.

I remember seeing packages that offer a GUI to edit Matplotlib plots. Found one (not sure it’s the one I was looking for) that does that: plt-editor-tool. It has an accompanying user guide in PDF from which I’ve made this screenshot:

The advantage of Explorer is that it’s general enough that it can be understood as data explorer and hvPlot’s API explorer. I also quite like, or even prefer, hvPlotEditor (hvplot.editor(df)).

[maximlt]

Yes it will eventually be documented! Here’s how you could start experimenting with it, by running this in a notebook:

import pandas as pd
import numpy as np

from hvplot.ui import hvPlotExplorer
import hvplot.pandas

df = pd.DataFrame(np.random.rand(20, 2), columns=['a', 'b'])
hvPlotExplorer(df)