7 Your Python environment

8 Your Python environment

The book is Python only. This chapter sets the floor: a working interpreter, a way to manage packages, and a way to keep one project’s libraries from breaking another’s. It is the least exciting chapter in the book and one of the most important.

8.1 What you will get from this chapter

A Python install that runs the book’s scripts.
A package manager you can rely on.
One small habit — the virtual environment — that will save you a weekend later.

8.2 Pick a Python

You want Python 3.11 or newer. Get it from python.org if you have nothing, or from brew install python on macOS, or from your system package manager on Linux.

Confirm it works:

python --version          # macOS, Linux
python3 --version         # macOS sometimes
py --version              # Windows

You should see Python 3.11.x or higher. If a 2.7 shows up, your system is old enough to drink in most countries; install a current Python alongside.

8.3 Pick a package manager

A package manager installs and tracks the libraries your code uses. Three reasonable choices:

uv — new, fast, well-designed. Recommended for new projects. Install with pip install uv (yes, you install it with the thing it replaces) or via astral.sh/uv.
pip plus venv — built into Python. The conservative choice. Universally understood.
Anaconda — comes with a curated stack of scientific packages. Heavy. Fine if you already have it; not worth installing fresh.

This book’s examples assume uv or pip. The principles transfer.

8.4 The virtual environment habit

Each project gets its own virtual environment. A virtual environment is a folder that contains a Python plus a fresh set of libraries, isolated from everything else on your machine. This sounds boring; it prevents the most common multi-week debugging headaches in scientific Python.

With uv:

cd my-project
uv venv                 # creates .venv/ in this folder
uv pip install pandas matplotlib statsmodels

With pip and venv:

cd my-project
python -m venv .venv
source .venv/bin/activate     # macOS/Linux
.venv\Scripts\activate        # Windows
pip install pandas matplotlib statsmodels

The activation step changes which Python a fresh terminal uses. You will forget it sometimes; you will reactivate. That is normal.

Add .venv/ to your .gitignore (we already did, in the previous chapter).

8.5 The libraries this book uses

A short list, installed once per project:

pandas numpy matplotlib seaborn statsmodels
requests pydantic python-dotenv
anthropic openai     # one of these, for Part VI

Save the exact list to a file so others can reproduce it:

pip freeze > requirements.txt          # or: uv pip freeze > requirements.txt

Commit requirements.txt. Six months from now you will still be able to recreate the environment with pip install -r requirements.txt.

8.6 The verify script

Save the following as verify.py in any folder. Run it.

import sys
import importlib

print(f"Python: {sys.version.split()[0]}")
for pkg in ["pandas", "numpy", "matplotlib", "statsmodels"]:
    try:
        m = importlib.import_module(pkg)
        v = getattr(m, "__version__", "unknown")
        print(f"  {pkg:<12} {v}")
    except ImportError:
        print(f"  {pkg:<12} NOT INSTALLED")

If the four core packages all print versions, you are ready. If not, install whichever is missing and try again.

8.7 A note for textbook readers

If you are reading Békés–Kézdi alongside, the textbook’s coding appendix walks through the same install in more detail and ties it to the textbook’s case-study data. That is the better introduction; treat this chapter as the checklist.

8.8 Where AI helps · Where AI bluffs

Helps. Diagnosing a cryptic install error. Translating a pip command into a uv command. Suggesting a minimal requirements.txt for a specific task.

Bluffs. Recommending a package version that does not exist. Telling you to pip install something that has been renamed. If you get an error like “could not find version X for package Y,” that is a real signal — read what pip actually offers, do not let the AI guess.

8.9 Keep this with you, not the AI

The package manager you pick. Pick one and stick with it across your projects.
Whether to commit requirements.txt (yes, almost always) and what to put in it.
The Python version your project targets. Decide once; mention it in your README.

8.10 Try this

In a fresh folder, create a virtual environment, activate it, install pandas, run a one-line script that imports pandas and prints its version. Total time: under three minutes. You now have a clean place to work for the next chapter.

8.11 AI and me

How did AI support me here?
How did AI fail me?
How did AI extend me?

8.12 Where to go next

Your AI account — the last setup step before we get to the interesting parts.

--- title: "Your Python environment" --- # Your Python environment The book is Python only. This chapter sets the floor: a working interpreter, a way to manage packages, and a way to keep one project's libraries from breaking another's. It is the least exciting chapter in the book and one of the most important. ## What you will get from this chapter - A Python install that runs the book's scripts. - A package manager you can rely on. - One small habit — the virtual environment — that will save you a weekend later. ## Pick a Python You want **Python 3.11 or newer**. Get it from [python.org](https://www.python.org/downloads/) if you have nothing, or from `brew install python` on macOS, or from your system package manager on Linux. Confirm it works: ``` python --version # macOS, Linux python3 --version # macOS sometimes py --version # Windows ``` You should see `Python 3.11.x` or higher. If a `2.7` shows up, your system is old enough to drink in most countries; install a current Python alongside. ## Pick a package manager A package manager installs and tracks the libraries your code uses. Three reasonable choices: - **`uv`** — new, fast, well-designed. Recommended for new projects. Install with `pip install uv` (yes, you install it with the thing it replaces) or via [astral.sh/uv](https://docs.astral.sh/uv/). - **`pip` plus `venv`** — built into Python. The conservative choice. Universally understood. - **Anaconda** — comes with a curated stack of scientific packages. Heavy. Fine if you already have it; not worth installing fresh. This book's examples assume `uv` or `pip`. The principles transfer. ## The virtual environment habit Each project gets its own virtual environment. A virtual environment is a folder that contains a Python plus a fresh set of libraries, isolated from everything else on your machine. This sounds boring; it prevents the most common multi-week debugging headaches in scientific Python. With `uv`: ``` cd my-project uv venv # creates .venv/ in this folder uv pip install pandas matplotlib statsmodels ``` With `pip` and `venv`: ``` cd my-project python -m venv .venv source .venv/bin/activate # macOS/Linux .venv\Scripts\activate # Windows pip install pandas matplotlib statsmodels ``` The activation step changes which Python a fresh terminal uses. You will forget it sometimes; you will reactivate. That is normal. Add `.venv/` to your `.gitignore` (we already did, in [the previous chapter](00c-github.qmd)). ## The libraries this book uses A short list, installed once per project: ``` pandas numpy matplotlib seaborn statsmodels requests pydantic python-dotenv anthropic openai # one of these, for Part VI ``` Save the exact list to a file so others can reproduce it: ``` pip freeze > requirements.txt # or: uv pip freeze > requirements.txt ``` Commit `requirements.txt`. Six months from now you will still be able to recreate the environment with `pip install -r requirements.txt`. ## The verify script Save the following as `verify.py` in any folder. Run it. ```python import sys import importlib print(f"Python: {sys.version.split()[0]}") for pkg in ["pandas", "numpy", "matplotlib", "statsmodels"]: try: m = importlib.import_module(pkg) v = getattr(m, "__version__", "unknown") print(f" {pkg:<12} {v}") except ImportError: print(f" {pkg:<12} NOT INSTALLED") ``` If the four core packages all print versions, you are ready. If not, install whichever is missing and try again. ## A note for textbook readers If you are reading [Békés–Kézdi](https://gabors-data-analysis.com/getting-started) alongside, the textbook's coding appendix walks through the same install in more detail and ties it to the textbook's case-study data. That is the better introduction; treat this chapter as the checklist. ## Where AI helps · Where AI bluffs **Helps.** Diagnosing a cryptic install error. Translating a `pip` command into a `uv` command. Suggesting a minimal `requirements.txt` for a specific task. **Bluffs.** Recommending a package version that does not exist. Telling you to `pip install` something that has been renamed. If you get an error like *"could not find version X for package Y,"* that is a real signal — read what `pip` actually offers, do not let the AI guess. ## Keep this with you, not the AI - The package manager you pick. Pick one and stick with it across your projects. - Whether to commit `requirements.txt` (yes, almost always) and what to put in it. - The Python version your project targets. Decide once; mention it in your README. ## Try this In a fresh folder, create a virtual environment, activate it, install `pandas`, run a one-line script that imports pandas and prints its version. Total time: under three minutes. You now have a clean place to work for the next chapter. ## AI and me 1. How did AI **support** me here? 2. How did AI **fail** me? 3. How did AI **extend** me? ## Where to go next [Your AI account](00e-ai-account.qmd) — the last setup step before we get to the interesting parts.