Sharing Jupyter Notebooks via MyBinder

Last Update: 9 Dec 2020

Learn how to publish Jupyter Notebooks on MyBinder. Launch an example of a shared Jupyter Notebook in binder here:



Analysis pipelines written in Python can be shared in an interactive document called a Jupyter Notebook. Notebook sections are usually one of two types:

  • Markdown - text/documentation of your analysis
  • Code - Python code chunks that run the analysis


Providing one document (notebook) of executable code and documentation eases reproducible research, where other scientists can verify, understand, and possibly extend your analysis.

To demonstrate these benefits, this tutorial will walk you through:

  • creating a minimal Jupyter Notebook
  • pushing it to a public github
  • and linking the notebook to a to be shared

We will assume you have a GitHub account and are using conda to manage python environments.

More Information

Step 1: GitHub

We will assume you have a GitHub account. If not, please register a free account.

  1. Sign in to Github -
  2. Create a jupyter_binder repository

Jupyter Repo

  1. Clone the repo, to create a copy on your local compuer.

git clone

Click on Code and the clipboard icon, to copy the git repo path. That way you can type git clone and paste the git repo path.

  git clone
  cd jupyter_binder

We will be adding a basic notebook to this folder, and pushing (publishing) it to GitHub.

Step 2: Create a minimal Jupyter Notebook

Creating and hosting a local Jupyter Notebook requires the jupyter package.

  • Use your favorite Text Editor and copy the following file, saving it as a environment.yml file. Notice how we’ve named the environment ntbk_env.

    name: ntbk_env
      - conda-forge
      - bioconda
      - defaults
      - python=3.8        #<= python version
      - jupyter           #<= list any python packages here
  • Create the python environment. From terminal, we can use conda env create to create a python environment. The -f flag means we are reading in dependencies from a file environment.yml. Creating an environment may take several minutes.

    conda env create -f environment.yml
  • Activate the environment.

    conda env list             # List all created environments
    #> conda environments:
    #> ntbk_env                 /Users/User_Name/miniconda/envs/ntbk_env
    conda activate ntbk_env    # Start the ntbk_env
    conda deactivate           # Stop the ntbk_env
  • Start the local Jupyter Notebook

    cd jupyter_binder/         # Navigate to git repo
    mkdir tutorials            # Create a folder
    conda activate ntbk_env    # Start ntbk_env
    jupyter notebook           # Starts notebook, should open a browser window
  • Create a Jupyter Notebook via New/Python 3

    Jupyter Local

  • We will try to recreate the Jupyter Notebook shown in the figure.


    For the moment, create one Markdown section and paste the following. It doesn’t matter what it says right now since we are testing the system out.

    ## Hello Jupyter Notebook
    This is a quick introduction to Jupyter Notebooks.
    * Markdown
    * Code

    Hit the Run button to render the markdown.

    Now let’s create a Code section and paste the following.

    print("Hello World")

    Hit the Run button and the words “Hello World” should show up under the Code section. Now save this Notebook (ipynb file) by going to File/Save As. Put the notebook in a tutorials folder.

    At this point, your jupyter_binder folder should look similar to:

      |_ tutorials/
      |    |_ Intro_to_Jupyter_Notebooks.ipynb
      |_ environment.yml
  • Save and commit to GitHub

    git add tutorials/Intro_to_Jupyter_Notebooks.ipynb
    git add environment.yml
    git commit -m "minimal notebook" tutorials/Intro_to_Jupyter_Notebooks.ipynb environment.yml
    git push origin main

    Check GitHub to make sure the new files are online.

  • Visit
  • Input the github name and any subfolder if ipynb files are in a subfolder. Note: Instead of test_binder, your repo will be called jupyter_binder.

My Binder

Click Launch and you should see your jupyter notebook in the browser. Congratulations, your binder is published!

Add the “Launch Binder” badge

The last thing is to add a badge that launches the binder. Notice in the above image, before you hit launch, there’s a “Copy the text, then paste into your README…”, expand that section and copy the badge code…which will look similar to:


Replace j23414 with your GitHub username, and test_binder with jupyter_binder. Then when you add it to the, you should have a working binder launch badge.


Congratulations! You have completed the Jupyter Notebook + Binder Tutorial.

Your Turn to Explore

From here on you are free to explore modifying the jupyter notebook. We recommend looking at some tutorials in Markdown and Python.

Note: About Python Module Dependencies

Remember that environment.yml file we used earlier? MyBinder will look for an environment.yml or a requirements.txt to find a list of python modules to install and make available in the notebook. This file needs to be at the base of the github repo (can’t find it inside of subfolder tutorials).

For example, if you want to use the Python package numpy then the environment.yml should look like:

  name: ntbk_env
    - conda-forge
    - bioconda
    - defaults
    - python=3.8        
    - jupyter           
    - numpy        #<= new numpy package

This installs numpy into the docker container, and you’ll be able to load the numpy package in a Code section:

import numpy as np

Together, the Jupyter Notebook + MyBinder + environment.yml provides a way to sharing a reproducible Python analysis pipeline.