I just recently decided to try R to analyze some data ive been generating.
I've always looked down on it as a "worse & less flexible python", but boy was I wrong. You can theoretically do all the same things in python as in R, but the R experience is so much smoother. 20 lines of setting up stuff in python is reduced to like 10 for the R equivalent.
ggplot2 is super intuitive compared to matplotlib. The inbuilt dataframes + tibbles are just as good if not better than anything pandas does. Besides all that, the libraries R has seem to be much higher in quality and in number, than python has, for data-science related stuff.
For my input, I think R (in particular ggplot2) is amazing for static graphs and rapid prototyping. When used with something like adobe illustrator for polishing, you can make some beautiful newsroom publication quality charts as well (see BBC).
However, JavaScript is the language of the web, and that dependency will never be removed from creating high-quality interactive visualizations. (Sure, you can write wrappers in R/Python, but you’ll still need to know the JavaScript under-the-hood to do anything bespoke/custom).
Don't disagree but you can get pretty far before you need to drop into js. Shiny [0] is one of the nicest UI frameworks I've come across. Period. I'm not a front end developer and it's always a mental struggle to get started with something that just looks nice. Shiny does - straight out the box. For interactive apps (e.g. dashboards) it's really quite compelling.
For non-interactive content, R Markdown is also pretty good. At the risk of being cast out as a heretic, I prefer Rmd to jupyter for report style docs (the kind of data science thing jupyter is meant for). Rmd isn't limited to R; it has multiple back ends including Python (which is also supported in the RStudio IDE).
So again: not disagreeing. If you need to get the polish that e.g. D3 can give, you'll need to get dirty with JS. But you can get a pretty long way before that.
BTW: I have no affiliation with RStudio the company, just a very happy user of their offerings (paid and open source).
ggplot2 is super intuitive compared to matplotlib. The inbuilt dataframes + tibbles are just as good if not better than anything pandas does. Besides all that, the libraries R has seem to be much higher in quality and in number, than python has, for data-science related stuff.