3.9 Jupyter Notebooks

Jupyter (formerly IPython Notebook) is an open-source project that lets you easily combine Markdown text and executable Python source code on one canvas called a notebook. It provides an interactive environment where you can write and execute code, visualize data, and document your analysis all in one place.

It sounds like R Markdown.

• One distinction is that R Markdown runs code in R console, while Jupyter notebook runs code in its own kernel.

• Jupyter can provide individual environments for each notebook, making it easier to manage dependencies and avoid conflicts between different projects. → But probably, it’s easier just to use the system R kernel as you don’t need to install packages in each environment.

  • You can use conda install to manage packages in each independent environment.

• Magic commands

  This can be used to run code in different languages within the same notebook. For example, you can use the %%R magic command to run R code in a Python notebook, or the %%python magic command to run Python code in an R notebook.

  There are many other magic commands available, such as %%bash to run shell commands, %%time to time the execution of a cell, and %%cd to change the current working directory.

• Online services

  You may have web-based access to Jupyter notebooks via services such as Google Colab, JupyterHub and Open OnDemand.

• You can convert Jupyter notebooks to Quarto qmd format:

  quarto convert demo-python.ipynb

  This is nice in part because qmd (like Rmd) is more easily handled by version control and with shell commands than the JSON format of .ipynb files.

Tutorial:

• IRkernel Install

• IRkernel CRAN Documentation

• Jupyter And R Markdown: Notebooks With R

• A simple example of Jupyter Notebook in R

• Jupyter Tutorial

• repr options

  repr is a package that provides rich representations for R objects in Jupyter notebooks. It allows you to customize how R objects are displayed in the notebook, including options for controlling the size and format of plots, tables, and other output.

  repr CRAN Documentation

1. Install IRkernel

   install.packages("IRkernel")
   IRkernel::installspec() 

   • Note this needs to be run in the R terminal or Radian (running in RStudio does not work).
   • IRkernel::installspec() makes the R kernel available to Jupyter. It lets Jupyter to use the default system R kernel instead of the R provided by Anaconda.

2. Reload window. Open command palette and type “Jupyter: Create New Blank Notebook” to create a new Jupyter notebook.

3. Click on the button right below ellipsis in upper right corner to choose kernel

   • Select Jupyter Kernels > R to use the R kernel.

     It should look something like this: