Jupyter & Colab Glossary
From notebooks and cells to JupyterHub and reproducibility tooling
- Jupyter Notebook - A web-based interactive document mixing executable code, output, equations, and prose.
- Cell - The basic execution unit of a notebook: a block of code or markdown run independently.
- Code Cell - A cell holding source code (Python by default) run by the active kernel.
- Markdown Cell - A cell holding formatted text in Markdown - headings, lists, links, equations, images.
- Kernel - The language-specific backend process that executes your code and holds variable state.
- Output - Whatever the kernel sends back after running a cell: text, tables, images, HTML, errors.
- .ipynb File Format - The on-disk JSON format for a Jupyter notebook: cells, outputs, metadata, version.
- JupyterLab - The next-generation web IDE for Jupyter - notebooks + file browser, terminal, editor, extensions.
- Google Colab - Google's free cloud-hosted Jupyter environment with managed Python, GPU, and TPU runtimes.
- Kaggle Notebooks - Free cloud Jupyter on Kaggle with GPU, pre-mounted datasets, and competition integration.
- Magic Command - A special IPython instruction prefixed with % (line) or %% (cell) beyond plain Python.
- Shift+Enter - The keyboard shortcut to run the current cell and advance to the next.
- Notebook Server - The local web server that serves your notebooks at http://localhost:8888.
- Anaconda - A Python distribution bundling conda, Jupyter, and ~250 data-science packages.
- Virtual Environment - An isolated Python install letting each project use its own package versions.
- pip - Python's standard package installer - fetches from PyPI and resolves dependencies.
- GPU Runtime - A notebook session backed by an NVIDIA GPU for fast deep-learning training.
- %matplotlib inline - A Jupyter magic that renders matplotlib plots as static PNG images inline below the cell.