does anyone have experience with compatibility between R-Shiny and Holoviews/Panel? My knowledge about the backend infrastructure is too poor for estimating the possible ways.
I remember seeing a R-Bokeh package in the Anaconda Library, but does it mean that the output from holoviews with the bokeh backend can be merged with a project written in R?
That’s a lot of questions all together, so I’ll try to answer each one:
- RBokeh is an R interface to BokehJS. Bokeh is made up of two big parts, BokehJS and Bokeh Python. Users program using Bokeh Python, and Bokeh Python communicates with BokehJS to actually draw the plots. If you want to use R to write a Bokeh plot, you can use RBokeh, but (a) RBokeh isn’t maintained nearly as well as Bokeh Python, so you won’t find everything working quite as smoothly, and (b) RBokeh doesn’t help you use Bokeh from Panel, because Panel is in Python and is already set up to use Bokeh Python. So I don’t think RBokeh is useful for integrating R and HoloViews/Panel/HoloViz.
- Shiny is a package that is competitive with Panel, and I don’t think there’s any useful way to combine them. Like Panel, Shiny lets you put widgets together and make dashboards, but it does it using R rather than Python. If you want to use R to make dashboards, use Shiny; if you want to use Python to make dashboards, use Panel.
- Although it’s Shiny isn’t useful for Panel, Panel can use plots from R. E.g. here is an example of using an R ggplot figure using rpy2: https://anaconda.org/jbednar/panel_gapminders/notebook R generates the plot, then Panel puts it into a dashboard. In this way you can use analysis tools and plotting libraries from R, but you still build the dashboard in Python.
Hope that helps!