Jupyter Notebook

Overview

Jupyter Notebook lets you write and run Python in your browser: live code, notes, and charts in one document. It's built on the Jupyter team's scipy image, so the scientific stack (NumPy, pandas, Matplotlib, scikit-learn, and more) is preinstalled and ready to import, with nothing to set up before your first chart.

Features

• Live Python code cells alongside notes and rendered charts in one document
• The scipy stack (NumPy, pandas, Matplotlib, scikit-learn, and more) preinstalled and ready to import
• Notebooks saved as ordinary files in your Documents folder, not buried in app state
• Install extra Python packages yourself with pip install --user

Setting up

On the login screen, enter the one-time setup token and choose a password. That password is how you sign in from now on, and it works straight away:

• Setup token: malmo-setup
• Password: choose your own on this screen

Then restart the app once: open its page in Settings and use Stop, then Start. Your password already works without this; the restart's only job is to switch off the malmo-setup setup token, which otherwise stays valid as a second way in until the app is next restarted (any later update or reboot also closes it).

To change your password later, open a Terminal inside Jupyter, run jupyter server password, and restart the app the same way.

Good to know

• Notebooks live in your Documents folder, in a subfolder you choose when you install ("Notebooks" by default). They are ordinary files on your drive: they stay yours, and they stay put if you uninstall the app.
• Install it just for yourself. Jupyter has no user accounts of its own, so a single household copy would mean everyone shares one password and one terminal into the same Notebooks folder. Install a personal copy per person instead.
• pip install --user is the supported way to add packages, and it survives restarts and updates. conda install, mamba install, and a plain pip install (without --user) do not work here: the app runs as a locked-down user that cannot change the preinstalled environment. The shipped scipy stack plus pip install --user covers everyday data work.
• No sudo or apt install inside the sandbox.
• No GPU acceleration, so heavy ML training is out of scope. This is built for interactive, CPU-based data work.