Overview

Teaching: 10 min
Exercises: 10 min

Questions

• What are Jupyter Notebooks?

• What is JupyterHub?

Objectives

• Learn how to share executable code that you’ve written

Jupyter Notebook

• A Jupyter Notebook App is a web application for running and sharing notebook documents.
• A notebook document can contain code and other elements that are executable via a kernel. It can also contain rich text and figures.
• A kernel executes the code in a notebook document. A kernel can be an environment (e.g. a conda environment)
• A Jupyter Notebook can help encourage collaboration, and reduces the need to “reinvent the wheel”

Installing the Single User Jupyter Notebook server

We will use some of the skills learned in the Github and conda tutorial to clone a version of the tutorial contents (https://github.com/geohackweek/tutorial_contents.git) on to your local machine. Then we will use conda to install a sample environment.yml file and run notebooks locally.

10 minutes git and conda challenge

1. Check if you have jupyter notebook installed (open terminal, run jupyter notebook)
2. Git fork (https://github.com/geohackweek/tutorial_contents.git)
3. Git clone your OWN fork of the repo
4. conda install the sample environment file from the raster tutorial
5. Launch Jupyter Notebook from inside the git repository
6. Test running one of the notebooks

What is Jupyterhub

Jupyterhub:

• Manages authentication
• Spawns Jupyter Notebook servers on-demand
• Each Jupyter Notebook server is self-contained

All notebooks have access to the same set of packages and libraries as specified in the environment file.

Which means…

Jupyterhub is a way to give a standardized Jupyter Notebook server to each person in a group of people

JupyterHub Access

The JupyterHub for GeohackWeek is accessible at: https://jupyterhub.geohackweek.org

You will need a Github account for authentication.

Key Points

• The cloud provides on-demand access to infinite computational resources

• Resources need to be carefully managed, because charges are usually tied to how long resources are held

• The cloud is great for bursty, high-volume computing, or for some small services you might want to run

---