Jupyter Notebook R Kernel for Microsoft R Open

Have you ever tried to add an R Kernel to Jupyter Notebook, having installed R Tools for Visual Studio? If tried, you know that this is not a piece of cake. I’ll show you how to do this.

The Jupyter Notebook is a web application in which you can create and share documents with code, text, and visualizations. Fresh Jupyter Notebook contains only one Python kernel. If you are using R language, installed from CRAN, you can easily add it to the Jupyter, following instructions on the page Native R kernel for Jupyter.

But that’s another matter if you installed Microsoft R Open (MRO). In fact, it could be set up separately or as a part of R Tools for Visual Studio (RTVS). Above-mentioned recipe won’t work for MRO.

First, let’s install some packages to the MRO library folder. On my machine, this is C:\Program Files\Microsoft\R Client\R_SERVER\library

Open system menu. Find Microsoft R Client > RGui, and run it as Administrator. Copy the following code to the R Console and run it:

lib.path <- 'C:/Program Files/Microsoft/R Client/R_SERVER/library'
install.packages('devtools', dependencies = TRUE, lib = lib.path)
install.packages(c('repr', 'IRdisplay', 'evaluate', 'crayon', 'pbdZMQ', 'uuid', 'digest'), lib = lib.path)

MRO on my machine has version 3.3.2, and I would try to install R kernel with the command:

IRkernel::installspec(name = 'ir33', displayname = 'R 3.3.2')

The result will be disappointing:

In fact, the error message gives us a clue: jupyter.exe cannot be found. Find a folder where the Anaconda is installed. On my machine, it is in the folder C:\Program Files\Anaconda3. Add folder C:\Program Files\Anaconda3\Scripts to the PATH environment variable of your account:

Close and open RGui and try again:

IRkernel::installspec(name = 'ir33', displayname = 'R 3.3.2')

Success! Now open system menu, and run Anaconda3 > Jupyter Notebook. R kernel should be in the list of notebooks:


Now you can remove C:\Program Files\Anaconda3\Scripts from the PATH variable if you don’t need it anymore.

Create a new R notebook and check that it is working now:


Added kernels could be found in the AppData\Roaming\jupyter\kernels folder like this: C:\Users\Eugene\AppData\Roaming\jupyter\kernels

If you don’t need a kernel, just remove it from this folder.