How does the Plotly backend for pandas work?

I just realized that one can set Plotly to be the backend of pandas:

However, according to Extending pandas โ€” pandas 1.6.0.dev0+268.ge25aa9d313 documentation, this is done by adding a pandas_plotting_backends entry point. And yet, when I display my entry points, I donโ€™t see a Plotly one:

In [13]: eps = entry_points()

In [14]: eps["pandas_plotting_backends"]
(EntryPoint(name='holoviews', value='hvplot:plotting', group='pandas_plotting_backends'),
 EntryPoint(name='matplotlib', value='pandas:plotting._matplotlib', group='pandas_plotting_backends'))

In [15]: import pkg_resources
    ...: for entry_point in pkg_resources.iter_entry_points('pandas_plotting_backends'):
    ...:     print(entry_point)
matplotlib = pandas:plotting._matplotlib
holoviews = hvplot:plotting

Then how does the backend work?

Hello @astrojuanlu

Take a look this post. Maybe can help you.

1 Like

Hi @SamuelVT , thanks for sharing the blog post, itโ€™s very insightful. However, I donโ€™t think it addresses my question of how does the underlying mechanism work.