[package proposal] vaex-plotly

[JovanVeljanoski]

This PR introduces a package proposal vaex-plotly which allows one to easily make figures (scatter plots, histograms, and heat maps) via Plotly.

Vaex is doing all the heavy lifting (aggregation, calculation of statistics etc..), and Plotly is providing the plotting backend. The returned figures are of type FigureWidget, making them suitable for dashboards and interactive functionality.

The methods are added as a DataFrameAccessor and can simply be accessed as

df.plotly.histogram
df.plotly.heatmap
df.plotly.scatter

which are based on plot1d, plot, and scatter from the vaex-viz package respectively. Note that not all functionality is ported.

Please check out the vaex-plotly-example.ipynb notebook in the examples folder.

Here are some examples on how the figures with Plotly look like.

[maartenbreddels]

The CI fails because vaex-plotly is not being installed.

[JovanVeljanoski]

Hi @maartenbreddels Yeah I was thinking about the interactive part as well, but I though of having a first iteration like this and build it more after you see the current state. I would propose to have a keyword, which would make the zooming/panning re-calculate the gird, or keep it fixed (as it is now). Would you be happy with this kind of solution?

Ah yeah, i was suspecting that the installation is most probably not set-up correctly.. so any help on this would be much appreciated.

[maartenbreddels]

Ah yeah, i was suspecting that the installation is most probably not set-up correctly.. so any help on this would be much appreciated.

see .travis.yml and appveyor file (it should be clear when you see them.

[maartenbreddels]

You've made it py3 only, the sourcecode, so I think you have to not install it even (or remove the _'s)