jupyter
npx machina-cli add skill G1Joshi/Agent-Skills/jupyter --openclawJupyter
Jupyter is the de facto standard for interactive data science. v7 (2025) of the Notebook is built on JupyterLab components, offering a modern, extensible experience.
When to Use
- Exploratory Data Analysis (EDA): Plotting data inline (
matplotlib). - Education: Teaching code with markdown explanations.
- Prototyping: Testing snippets before moving to a script.
Core Concepts
Kernels
The computation engine (IPython, IJulia).
Cells
Code cells (executed) vs Markdown cells (documentation).
Magic Commands
%timeit, !pip install.
Best Practices (2025)
Do:
- Use JupyterLab: The richer, multi-tab interface is standard.
- Use
nbdev: If you want to build libraries from notebooks. - Use Version Control: Use
jupytextto pair notebooks with.pyfiles for git diffs.
Don't:
- Don't store secrets: Clear output before committing.
References
Source
git clone https://github.com/G1Joshi/Agent-Skills/blob/main/skills/ai-ml/jupyter/SKILL.mdView on GitHub Overview
Jupyter is the de facto standard for interactive data science. JupyterLab v7 (2025) builds on a modern, extensible notebook interface, enabling rich data exploration workflows.
How This Skill Works
Jupyter uses kernels (e.g., IPython, IJulia) as the computation engine and distinguishes between code cells and Markdown cells. Users can run code inline, render plots, and annotate results with markdown, while magic commands like %timeit and !pip install extend functionality. For workflows, tools like nbdev and jupytext enable library development and version control directly from notebooks.
When to Use It
- Exploratory Data Analysis (EDA) with inline plotting (matplotlib).
- Education: teaching code with markdown explanations.
- Prototyping: testing snippets before moving to a script.
- Interactive experimentation and visualization to gain quick insights.
- Library or pipeline development from notebooks using nbdev and jupytext for version control.
Quick Start
- Step 1: Install JupyterLab and start it (e.g., pip install jupyterlab; jupyter lab).
- Step 2: Create a notebook, select an IPython kernel, and mix code cells with Markdown; run cells to explore data.
- Step 3: Save and optionally pair notebooks with .py using jupytext for git diffs (consider nbdev for libraries).
Best Practices
- Use JupyterLab for the richer, multi-tab interface.
- Use nbdev to build libraries from notebooks.
- Use version control with jupytext to pair notebooks with .py files for git diffs.
- Don't store secrets; clear output before committing.
- Keep notebooks modular and well-documented to aid reuse and collaboration.
Example Use Cases
- EDA notebook with inline matplotlib plots to explore data distributions.
- Educational notebook combining code cells with Markdown explanations for a lesson.
- Prototype a data transformation or model snippet before integrating into scripts.
- Develop a library workflow from notebooks using nbdev.
- Version-control notebooks by pairing them with .py files via jupytext for clean diffs.
