Skip to main content

Trying out Positron for R use

Positron seems nice

Recently the people behind Rstudio/Posit have announced the release of Positron, a new IDE for R. I guess this new IDE address the will of Posit to move more away from the R-only world, as they already did with the company rebranding and with the release of Quarto.

Positron is basically Visual Studio Code but tailored for R and Python. I moved away from RStudio a while ago because I found it too sluggish, oldish and not as flexible as I wanted. I am using VSCodium so I am definitely non against using that as a base for something new.

I did try Positron for a while (I was playing around with the kairos package) and I am definitely liking it. There are still things missing (e.g. RStudio addins) but it kind of feels as a nice IDE without needing to setup all the R extensions for VSCode+Radian console. Everything is very fast, plots visualization is way better than both RStudio and VSCode (this is the most painful part of not using RStudio), and viewing data.frame/matrix is a great experience.

Positron IDE with code and plot visualization
Positron IDE - I was testing out the kairos package by N. Frerebeau

The View() function brings up a nice viewer with the df just as RStudio/VSCode, but with a new pane that provides a list of columns and some summary data about each column. I find the graph visualization a bit confusing (the percentage is the amount of missing data), but the whole idea is quite cool.

Positron IDE View with the new data frame viewer
The new dataframe viewer, again using the kairos package as example

All in all, it seems like a nice tool, I would have preferred for them to use the AGPL license (open-source) just as with RStudio, instead of the Elastic license (source-available). But I am a dreamer.

A nice(r) presentation of Posiron with some actual useful tips and tricks was recently published by Andrew Heiss, definitely worth a read!

Addendum#

I have no real sympathy for Posit as a company, due to some past stuff about collaborating with Palantir, and the weird PR answers that followed. I believe a company that wants to be seen a good open-source citizen should hold up to better standards, especially when choosing partners (even if it doesn't impact users directly).

I also kind of disagree with many dev/user resources that silos people towards using the company's tools such as RStudio or the company's packages such as the Tidyverse. R is accessible to everyone regardless of the tools/libraries, and resources should reflect that.

Reply by email
Back to top