Statistical Rethinking Practice

A collaborative learning environment for working through Statistical Rethinking exercises in Python using NumPyro.

What's This Repository?

This is a shared workspace where group members can:

• Work through Statistical Rethinking problems from the book
• Use a consistent, reproducible Python environment via Nix
• Share their solutions with others (optional)
• Reference different approaches to the same problems

Repository Structure

statistical-rethinking-practice/
├── shell.nix                      # Nix environment configuration
├── starter_template.ipynb         # Template to copy for each chapter
├── reference_solutions/           # Canonical solutions from external repos
│   ├── chapter_02_solutions.ipynb
│   ├── chapter_02_numpyro_guide.ipynb
│   ├── chapter_03_solutions.ipynb
│   └── chapter_03_numpyro_guide.ipynb
├── chapter_02/                    # Each chapter gets its own folder
│   ├── alice.ipynb               # Individual member solutions
│   └── bob.ipynb
├── chapter_03/
│   └── carol.ipynb
└── ...

When you work on a chapter, you'll create your own notebook from the starter template (such as, chapter_02/yourname.ipynb) in a chapter-specific folder. You can optionally share your solutions to let everyone see different approaches to the same problems.

---

First Time Setup

You only need to do this once.

Step 1: Open Terminal

On macOS:

1. Search for "Terminal" in Spotlight
2. A window will open with a command prompt (something like yourname@computername ~ %)

Step 2: Ensure You Have Nix Installed

To check if you have Nix installed, open Terminal and run:

nix --version

If you see a version number, skip to Step 1. If you get "command not found", install Nix:

curl -L https://nixos.org/nix/install | sh

This will take 5-10 minutes. When it finishes:

1. Close Terminal
2. Open a new Terminal
3. Verify installation: nix --version

Step 3: Clone the Repository And Go Into It

If you haven't already cloned this repository:

git clone <repository-url>
cd statistical-rethinking-practice

To confirm you're in the right place, type:

ls

You should see files like shell.nix, README.md, and starter_template.ipynb listed.

Step 4: Enter the Nix Shell

Type this command and press Enter:

nix-shell

What happens: The first time you run this, it will:

• Download Python 3.11
• Create a virtual environment (a .venv/ folder)
• Install Python packages: arviz, matplotlib, jax, numpyro, jupyter

This will take several minutes. You'll see a lot of text scrolling by. That's normal! Wait until you see:

==========================================
Statistical Rethinking environment ready!
Run 'jupyter notebook' to start.
==========================================

When you see your command prompt again (it might now have (nix-shell) or similar in front of it), the setup is complete.

Note: Every time you open a new Terminal window to work on this project, you'll need to:

1. Navigate to the project folder (cd /path/to/statistical-rethinking-practice)
2. Run nix-shell (it will be fast after the first time)

Step 5: Test Jupyter

Still in Terminal, type:

jupyter notebook

What happens: Your web browser will open automatically, showing the Jupyter interface with a list of files in the project.

To test everything works:

1. Click on starter_template.ipynb to open it
2. Click on the first code cell (the one with all the import statements)
3. Press Shift + Enter to run it
4. You should see "Environment ready!" below the cell

To stop Jupyter: Go back to Terminal and press Ctrl + C, then type y and press Enter.

Working on a New Chapter

This is what you'll do each time when starting a new chapter.

Step 1: Get Up to Date Information and Start Nix

cd /path/to/statistical-rethinking-practice
git pull
nix-shell

If git pull throws an error, you probably need to tell git about your local work first. You can do this with

git add .
git commit -m "finished chapter 2 problems"

Step 2: Create a Chapter Folder

If this is the first time anyone is working on this chapter, create a folder for it:

mkdir chapter_02

Replace 02 with whatever chapter number you're working on.

What this does: mkdir stands for "make directory" (directory = folder). This creates a new folder called chapter_02.

Step 3: Copy the Starter Template

Copy the template and rename it with your name:

cp starter_template.ipynb chapter_02/yourname.ipynb

What this does: cp stands for "copy". This copies starter_template.ipynb into the chapter_02 folder and names it yourname.ipynb.

Example:

cp starter_template.ipynb chapter_02/alice.ipynb

Step 4: Start Jupyter

jupyter notebook

Your browser will open with the Jupyter file browser.

Step 5: Open Your Notebook

1. In the Jupyter file browser, double-click the chapter_02 folder
2. Click on your notebook (e.g., alice.ipynb) to open it
3. Run the first cell (the setup cell with imports) by clicking on it and pressing Shift + Enter
4. You should see "Environment ready!"

Now you're ready to add your own cells and work through problems!

---

Working in Jupyter

Running Cells

• Run current cell and move to next: Shift + Enter
• Run current cell and stay on it: Ctrl + Enter

Adding Cells

There are two ways to add a new cell:

Option 1: Using the toolbar

• Click the + button in the toolbar at the top

Option 2: Using keyboard shortcuts (in command mode)

• Press B to insert a cell below
• Press A to insert a cell above

Command mode vs Edit mode:

• When you click inside a cell and can type, you're in "edit mode"
• Press Esc to enter "command mode" (cell has blue border)
• Press Enter to go back to edit mode

Cell Types

Code cells (for Python code):

• This is the default
• When you run a code cell, Python executes the code and shows any output below
• Press Y in command mode to convert a cell to code

Markdown cells (for notes and explanations):

• Used for writing text, explanations, or formatting
• Press M in command mode to convert a cell to markdown
• After writing, run the cell (Shift + Enter) to render the formatted text

Useful Jupyter Tips

• Save your work: Cmd + S (Jupyter auto-saves periodically too)
• Restart if things get stuck: In the menu, choose Kernel → Restart & Clear Output
• See all keyboard shortcuts: Press H in command mode

Referencing Solutions

Reference Solutions

Check the reference_solutions/ folder for solutions from external repositories.

Later chapters don't have reference solutions publicly available in python, so we'll need to use each other if we want to check our work.

Other Members' Solutions

To see how someone else approached a problem:

1. In Jupyter file browser, navigate to the chapter folder (e.g., chapter_02/)
2. Click on their notebook (e.g., bob.ipynb)
3. Look through their approach

This is great for learning different ways to solve the same problem!

To get an up to date version of other people's solutions, run

git pull

Sharing Your Solutions (Optional)

Want to share your work with the group? Here's how:

Step 1: Make Sure Jupyter is Saved

Save your notebook in Jupyter (Cmd + S), then you can close Jupyter (in Terminal: Ctrl + C, then y).

Step 2: Check What Changed

In Terminal, type:

git status

This shows you which files have changed. You should see your notebook listed under "Untracked files" or "Changes not staged for commit" in red.

Step 3: Stage Your Changes

git add chapter_02/yourname.ipynb

This tells git you want to include this file in your next save/commit.

You can also run

git add .

to tell git to add every local change.

Step 4: Commit Your Changes

git commit -m "Add my solutions for chapter 2"

This creates a save point (called a "commit") with a message describing what you did.

Step 5: Push to GitHub

git push

This uploads your commit to the shared repository so others can see it.

Note: Sharing is Completely Optional!

You can keep your work local if you prefer. Only push when you're comfortable sharing.

Common Workflows

Starting a Work Session

# 1. Open Terminal
# 2. Navigate to project
cd /path/to/statistical-rethinking-practice

# 3. Get any updates from the repo
git pull

# 4. Start nix shell
nix-shell

# 5. Start Jupyter
jupyter notebook

# 6. Open your notebook and start working!

Before You Finish

# 1. Save your work in Jupyter (Cmd + S)
# 2. Stop Jupyter (Ctrl + C in Terminal, then y)
# 3. Exit nix-shell (optional)
exit

# 4. If you want to share your work:
git add chapter_XX/yourname.ipynb
git commit -m "Description of what you did"
git push

Troubleshooting

"command not found: nix-shell"

Nix might not be installed. Check with the group about Nix installation on your machine.

"I got a Python error when running code"

This is part of learning! Common issues:

• NameError: You might need to run earlier cells first (especially the imports cell)
• SyntaxError: Check for typos in your code
• Try restarting the kernel

"Git says there's a merge conflict"

This happens if two people edited the same file. Since everyone works in their own notebook (yourname.ipynb), this should be rare. If it happens, ask for help!

"The .venv folder appeared - should I commit it?"

No! The .gitignore file already excludes it. This folder is created locally on each person's machine and shouldn't be shared.

Environment Details

What's Installed

• Python 3.11
• Jupyter Notebook - Interactive coding environment
• arviz - Bayesian analysis and visualization
• matplotlib - Plotting and graphics
• JAX - Numerical computing (like NumPy but faster)
• NumPyro - Probabilistic programming (for Bayesian models)

How the Environment Works

1. Nix provides Python 3.11 at the system level
2. Python's venv creates an isolated environment (the .venv/ folder)
3. pip installs Python packages into that environment
4. The shell hook (in shell.nix) automatically activates everything when you run nix-shell

This means everyone in the group has exactly the same Python version and packages - your code will work the same on everyone's machine!

Adding More Packages (Advanced)

If you need an additional Python package:

# While in nix-shell:
pip install package-name

It will install to your local .venv/ folder. To make it permanent for everyone, the shell.nix file would need to be updated.