Jupyter Notebook Basics

A Jupyter notebook is a browser-based editor where you mix code cells, markdown cells, and their outputs in one file. Run a cell, see the result right below, write an explanation, continue. It is the default tool for data science because it fits how exploration actually works.

The two main cell types

• Code cells: run Python (or R, Julia, etc.) and show output
• Markdown cells: formatted text with headings, links, LaTeX math

Getting started

# Install
pip install jupyterlab pandas matplotlib

# Launch
jupyter lab

# Or use cloud versions
# - Google Colab (colab.research.google.com)
# - Kaggle Notebooks
# - Hugging Face Spaces with Jupyter
# - GitHub Codespaces

The kernel is the key concept

A kernel is the running Python (or other) process that executes your cells. Variables persist across cells because they live in the kernel. This is why the order you run cells matters more than the order they appear in the notebook.

Must-know shortcuts

• Shift+Enter: run cell, go to next
• Ctrl+Enter (Cmd+Enter Mac): run cell, stay here
• A: insert cell above, B: insert below (command mode)
• M: convert to markdown, Y: convert to code
• 00: restart the kernel

Magic commands

# Time a cell
%%time
expensive_function()

# Inline plots
%matplotlib inline

# Run shell commands
!pip install seaborn
!ls -lh data/

# Inspect a variable
?pd.read_csv  # shows docstring

Common anti-patterns to avoid

• Giant cells with 200 lines of code (break them up)
• No markdown explaining the flow (readers drown)
• Leaving unused code in the middle of the notebook
• Hardcoding file paths that only work on your machine
• Not pinning library versions for reproducibility

The big idea: Jupyter is exploration married to narration. Treat your notebook as a story, not just a scratch pad, and your future self (and collaborators) will thank you.