Jupyter Notebooks

Interactive computing environment combining code, visualizations, and narrative text

Developer Tools & Best Practices Essential Core Tool Recommend Provisional Teaching

Our View

Recommend Provisional Teaching

Jupyter notebooks allow you to combine code, results, and explanations in a single document. This makes them perfect for exploratory data analysis, teaching, and creating reproducible research workflows. You can iterate quickly, visualize results inline, and document your thought process as you work.

Provisional positions are editorial proposals awaiting group review.

Teaching — Recorded in topic offerings; consult the dated event links for availability.

6 topics with recorded teaching · 15 past offerings · 2 upcoming offerings

Related themes: Scientific Computing

Common use cases

Exploratory data analysis and prototyping
Document analysis workflows with code and explanations
Share reproducible analyses with collaborators
Create tutorial materials and teaching content

Getting started

Jupyter (formerly IPython Notebook) is an open-source web application that allows you to create and share documents containing live code, equations, visualizations, and narrative text. It’s become the standard tool for interactive data science and research computing.

Why Jupyter?

Interactive: Write code, run it, see results immediately
Literate Programming: Mix code with markdown explanations
Visualization: Display plots and figures inline
Reproducible: Share notebooks with all code and outputs
Multi-Language: Supports Python, R, Julia, and more

Key Features

Code Cells

Execute Python code interactively:

import numpy as np
import matplotlib.pyplot as plt

# Generate data
x = np.linspace(0, 10, 100)
y = np.sin(x)

# Plot inline
plt.plot(x, y)
plt.show()

Markdown Cells

Document your analysis with rich text:

Headers, lists, bold, italic
LaTeX equations: $E = mc^2$
Links and images
Code blocks with syntax highlighting

Rich Output

Display diverse content types:

Plots and figures
DataFrames (pretty HTML tables)
Images and videos
Interactive widgets
HTML and JavaScript

Installation & Usage

Install Jupyter:

conda install jupyter
# or
pip install jupyter

Launch notebook server:

jupyter notebook
# or for JupyterLab (modern interface)
jupyter lab

Jupyter Lab vs Classic Notebooks

JupyterLab (recommended):

Modern interface with multiple panels
Integrated file browser and terminal
Extensions and themes
Better for complex projects

Classic Notebook:

Simpler, single-document interface
Lighter weight
Good for quick analyses

Common Workflows in Research

Exploratory Analysis

# Load data
import pandas as pd
data = pd.read_csv('experiment.csv')

# Quick exploration
data.head()
data.describe()

# Visualize
data.groupby('condition')['reaction_time'].hist(alpha=0.5)

Parameterized Notebooks

Use tools like Papermill to run notebooks with different parameters:

papermill analysis.ipynb output.ipynb -p subject_id S01

Publishing Results

Convert notebooks to:

HTML reports
PDF documents
Slideshows
Blog posts (with Hugo, Jekyll)

Best Practices

Restart & Run All: Ensure notebooks execute in order
Clear Outputs: Before committing to version control
Short Cells: Keep individual cells focused
Document: Explain what and why, not just how
Version Control: Commit .ipynb files to Git
Virtual Environments: Use separate environments per project

Tips

Use keyboard shortcuts (Esc for command mode, Enter for edit mode)
Learn magic commands: %matplotlib inline, %timeit, %%time
Use tab completion and ? for documentation
Split complex analyses across multiple notebooks
Use %load_ext autoreload for developing packages
Consider nbconvert for generating reports