My Docker Template for R


Kazuharu Yanagimoto


May 24, 2022

Docker is one of the best environments for reproductivity of the research. I use Docker for basically all my projects, but it took me about 2 years to get to a satisfactory environment setup. I would like to share it here as a template.

Quick Start

  1. Clone this repository
  2. Open it in VSCode and add Remote-Containers Extension
  3. From the command palette, choose “open folder in container”
  4. Open localhost:8787 in a browser
  5. Create a project for this project directory (by default, choose /home/rstudio/work)
  6. RUN renv::init() in the R console
  7. Set up a DVC environment
    1. Prepare a folder in Google Drive (and copy the folder code)

    2. Init DVC

      dvc init
      dvc remote add --default myremote gdrive://GDRIVE_FOLDER_CODE

Brief Explanations

In this section, I briefly comment on each file. I assume you are familiar with

I manage data files by DVC. But if you have another strategy for data management, ignore these parts.


FROM rocker/verse

RUN apt update && apt install -y gnupg openssh-client libpoppler-cpp-dev

RUN wget \ \
       -O /etc/apt/sources.list.d/dvc.list && \
    wget -qO - | apt-key add - && \
    apt update && \
    apt install -y dvc 

# R Packages
RUN R -e "install.packages(c('languageserver', 'renv'))"

# Rstudio Global Options
COPY --chown=rstudio:rstudio .config/rstudio/rstudio-prefs.json /home/rstudio/.config/rstudio/rstudio-prefs.json

R image

If you don’t have any preferences, I recommend you use one of the rocker image series of rocker-org:

  • rocker/rstudio
  • rocker/tidyverse
  • rocker/verse
  • rocker/geospatial

Because I currently write a paper by Rmarkdown, I usually use rocker/verse and use rocker/geospatial for spatial analysis.


I just follow the official guide for Install from repository. If your environment includes python, it is easier to install it via pip.

R packages

I use renv for dependency management. languageserver allows us to run a code in VSCode, but I don’t usually use it.

Rstudio Global Options

When Rstudio launches, it loads the ~/.config/rstudio/rstudio-prefs.json file for general settings. So I prepare /.config/rstudio/rstudio-prefs.json, and it allows me to use Rstudio with my favorite settings.

If you want to use your own settings, I recommend you

  1. Open Rstudio
  2. Set up by GUI (Tools -> Global Options)
  3. Copy ~/.config/rstudio/rstudio-prefs.json (in container)
  4. Paste it on /.config/rstudio/rstudio-prefs.json (in host)


version: '3'
            context: .
            - TZ=Europe/Madrid
            - DISABLE_AUTH=true
            - RENV_PATHS_CACHE=/home/rstudio/.renv/cache
            - .:/home/rstudio/work
            - $HOME/.renv:/home/rstudio/.renv/cache


You don’t need to type a password for each new session.

Cache for renv

By setting the RENV_PATH_CACHE and mounting the directory, you don’t need to re-install packages. In other words, you can share packages with other (docker) projects and don’t have to re-install all the packages when you rebuild a container.

Have fun!