Question: thoughts on a Python version?

[thomasballinger]

I'm thinking about making a Python wrapper to copy your library (unless you're thinking of making one?) like observable-jupyter-widget is for robservable, a la Altair for Vega.

What do you think about the API right now? Any suggestions for things to do differently?

[juba]

That certainly is a good idea !

Regarding the API, unfortunately I didn't really get any feedback from users except myself. The main differences between obsplot and Plot are listed in the introduction vignette, and for the moment I don't really regret any of them, but I miss a bit more usage and experience. For example, one thing that I find handy in obsplot is that you can specify your data source only once in the initialization call, and it will be passed automatically to marks and facet calls in the same plot, whereas in Plot you'll have to specify the data source for each one of them.

So unfortunately no really useful feedback to give you right now, but if you happen to have questions or comments while working on the Python wrapper, I'd be glad to discuss them with you.

[juba]

Just for the record, there is now a Python version : https://github.com/juba/pyobsplot