Adds binder abilities on top of the rocker/tidyverse images.

This approach works on any machine on which you have Docker installed.

The rocker/binder images can be run like any other docker image:

docker run -p 8888:8888 rocker/binder

Note that binder will run JuyterHub on port 8888 by default. The above command will print to the terminal (and the docker container log) the URL which includes a randomly generated token for secure login, so be sure to include that in the URL you paste into the browser.

Once inside JupyterHub, RStudio Server should be an option under the menu "New":

That should start you into an RStudio session (with no further login required).

This approach uses the public binder cloud and requires no installation

Add a Dockerfile with the following contents to the root of a GitHub repository.

FROM rocker/binder:3.4.2

Extend as necessary (see below). Then go to <beta.mybinder.org> and enter the https address of your repository to launch your instance on the public Binder infrastructure.

The 3.4.2 indicates the version of R requested. Use a specific version to ensure a consistent build environment or use latest to get the latest avialable version of R and R packages.

No additional lines are needed unless you want to install further libraries. In that case, you will need to first switch to root user to run apt-get, and then switch back to $NB_USER (rstudio) to keep JupterHub startup happy. For instance:

FROM rocker/binder:3.4.2

USER root
RUN apt-get update && apt-get -y install libv8-dev


USER ${NB_USER}

https://twitter.com/twitter/statuses/921165684188393472