How to install Jupyter on Windows

This guideline intends to help students install Jupyter Notebook on their Windows PC in an easy way. Usually it is easier to troubleshoot problems when a UNIX system is used. Windows users may wish to install Windows Subsystem for Linux for these purposes. A guideline on how to do this will be added later.

windows 10

Terminology

Python - Programming language created by Guido van Rossum. The syntax aims to be readable. Because a lot of science packages are available for Python, it is a popular language used by scientists in all kind of disciplines. The most notable of these packages are Pandas, Numpy and Scikit-learn.

python

Jupyter Notebook - This is one of the many IDE that are available for Python. Specific for Jupyter Notebooks is that you can create reports with it at ease, since you can combine code with text and images. These reports are stored as .ipynb files.

jupyter

Anaconda - As explained, Python has become one of the major programming languages because many packages are available. Anaconda allows to automate the installing of dependencies and the updating of packages. An additional advantage of Anaconda is that it will create its own virtual environment and hence will not mess with any system-installed Python version.

anaconda

Numpy - Python package that provides convenient functions that deal with (multi-dimensional) arrays.

numpy

Pandas - Python package that provides convenient functions that deal with tabular data. Largest difference with Numpy is that it allows to have indices on the axis.

pandas

Scikit-learn - Python package that allows to do machine learning.

scikit-learn

Step 1: Download and install Anaconda

You can download Anaconda at anaconda.com. Choose for the Python 3 version, since 2 is EOL. When downloading is done, you can run the installer and leave every option to the default value.

Step 2: Run Jupyter Notebook

Choose Start → Anaconda3 → Jupyter Notebook. A terminal will start. If there is some error in this terminal, you can try to run Jupyter Notebook with admin privileges by right clicking → More → Run as administrator. When everything went okay, the browser should automatically open at localhost:8888. If not, please open it yourself.

Step 3: Create a new jupyter notebook

Navigate to the folder where you want to create the notebook. Create the new notebook. A new file with the extension .ipynb should be added. Open this notebook by clicking on it.

create a new ipynb notebook

Step 4: Try

You can now run Python in the Jupyter Notebook! Play around a little. Try some additions (3+3) and check the result by executing the cell (can be done with shift enter). If at this stage you get a Kernel Error, close Jupyter Notebook (both the browser and terminal) and reopen it with administrator rights.

some example in jupyter notebook

Step 5: Check installed modules

Check which modules are installed. This can be done by executing the following code:

import pandas as pd # Load the pandas module.
import sklearn # Load the scikit-learn module.

If some package is not installed, you can open Start → Anaconda prompt (possibly with admin rights) and then type conda install .... The dots should be replaced by the package name. Jupyter Notebook should be reopened after installing new modules.

Step 6: Do your research!

Now you are suited with all tools necessary to do your research. As an example, you can download here some simple machine learning on sales data from the DSPM course. Save the .ipynb file to a file path to your choice. You can then start Jupyter Notebook and open the notebook file.

Download Example Jupyter Notebook on Machine Learning