How-to install Jupyter to virtualenv

Posted on Sun, 06 August 2017 in data (Last modified on Thu, 10 August 2017)

In this short how-to I will install Jupyter within a virtualenv running on Python 3.


Some prerequisites:

  1. I am running macOS 10.10+. At the time of writing, I am running macOS Sierra 10.12.2.
  2. I have Python 3.6 installed. At the time of writing, I am running Python 3.6.2 and pip 9.0.1 installed with Homebrew.

Update 10.8.2017:

  • The installation has been tested also on Windows 10 Pro, with Python 3.6.1 installed from

Create new virtualenv and install Jupyter

This can be done in multiple ways, I use my own handy tool to create a new project with virtualenv.

Now, that I have a new project set up, virtualenv active, and pip & setuptools upgraded to latest versions, I can move forward and install Jupyter and kernel.

pip install jupyter
python -m ipykernel install --user

With the installation complete, I can start the server with jupyter-notebook. This will open my default browser at http://localhost:8888/. Now I can create my first Jupyter notebook and start data developing.

Install Python packages

The environment is no good with no packages. I can install them the usual way by stopping the Jupyter server and installing them from command line. The other way is to install packages directly from Jupyter notebook by running this in a code cell (the ! enables the running of shell commands):

!pip install pandas

Now I have pandas installed within my virtualenv and I can inport the module within the workbook:

import pandas as pd

It’s important to manage the dependencies as always, so do not forget to build the requirements.txt with pip freeze >requirements.txt.

And that’s pretty much all there is to it. Happy Jupytering.