Jupyter Running – an introduction to JupyterLab and Notebooks

• 00:00 – Introduction
• 01:20 – Get JupyterLab up and running
  • 01:45 – github repo for this lecture
  • 02:13 – setup conda environment with dependencies
  • 04:10 – start up JupyterLab!
• 05:00 – the Jupyter command palette
• 05:35 – Why it’s a good idea to keep code OUT of your notebook
  • 06:30 – autoreload trick for keeping code in modules out of your notebook
  • 07:50 – use pycharm or any other IDE to edit those modules
• 09:20 – Jupyter Rising is a confusing band from the mid 2000s
• 11:00 – What exactly are Jupyter Notebooks and what are they good for?
• 12:40 – What is JupyterLab then?
  • 13:00 – JupyterLab also embeds an old-style notebook server
  • 13:37 – JupyterLab console that connects to a running notebook kernel
• 14:50 – Continue with demo

In this mini-lecture, we treat the following four main topics:

• Get JupyterLab up and running.
• The nature and role of Jupyter Notebooks and JupyterLab.
• Load and show data with a simple chart and a GIS visualization.
• The drawbacks of Jupyter Notebooks.

After following this mini-lecture, you will be able to:

• Explain the role of Jupyter Notebooks in the modern data analysis pipeline.
• Explain the risks of Jupyter Notebooks in the modern data analysis pipeline.
• Get your own JupyterLab up and running.
• Create new notebooks.
• Load datasets into Pandas dataframes.
• Build simple charts with those dataframes.

You will need an installed and confgured miniconda3 as explained in the vxuni mini-lecture Up and Running with Miniconda and PyCharm on MacOS.

Having followed the Conda Environments would be a bonus, but is not required.

I have created a github repo with the code and notebook for this mini-lecture.

Checkout the code it your working directory, for example:

mkdir ~/Documents/vxuni
cd ~/Documents/vxuni
git clone git@github.com:vxuni/jupyter_running.git
cd jupyter_running

Bring your vxuni conda environment up to date:

conda env update

Start JupyterLab:

jupyter lab

Double click on the jupyter_running.ipynb file in the File sidebar on the left.

In the video, I go through the basics of navigating and executing the wonderful world that is the Jupyter Notebook.

Jupyter Notebooks, previously known as IPython notebooks, are a web-based modality for doing literate programming.

In other words, you can combine documentation, runnable code, results such as charts, tables and other data, all in a single document.

Notebooks can be viewed online, and others can reproduce your results by running a notebook you made.

See A gallery of interesting Jupyter Notebooks by the folks behind Jupyter for an amazing list of instantly-viewable and reproducible notebooks.

JupyterLab is a new web-based software for scientific computation, with one of its primary operands being Jupyter Notebooks.

JupyterLab will eventually replaced Jupyter Notebook as the tool of choice for working with notebooks.

• Configure notebook.
  • Auto-reloading.
  • Notebook interactive charts.
• Embed YouTube video.
• Load and explore data.
• Build a chart.
• Build a more involved GIS visualization using folium.
• Demo the console connected to the notebook.

Notebooks are great for documentation, presentation and interactive work. However, they get in the way of:

• Using a full-powered IDE.
• Building well-structured software.
• Writing tests.
• Effective version control.

We have to keep these drawbacks in mind whilst drinking all of this refreshing Kool-Aid.

Personally, I try to mitigate these issues with the following guidelines:

• Keep notebooks as short as possible, for example limitng the notebook to a single experiment or hypothesis.
• Factor out code as you go along. That factored out code should be documented, version-controlled and tested when relevant.