Appendix A — Installation Guide

This appendix walks through setting up Python and an IDE (VS Code is recommended) on your own laptop for this course, so that every chapter in this book runs exactly as shown. The idea is to create a dedicated, self-contained virtual environment with a known-good set of package versions, rather than relying on whatever happens to already be installed on your machine.

A.1 Getting the course files

Before installing anything, get a local copy of the book’s repository. The notebooks refer to requirements.txt and the data/ folder by a relative path, so those only resolve correctly once you have the whole project folder, not just one downloaded notebook. Downloading a single notebook by itself, for example via right-click “Save link as” on one file, will not work: it won’t come with requirements.txt or data/ alongside it, and every pip install or pd.read_csv call in this book assumes they’re there.

The simplest way, no git required:

Go to the book’s GitHub repository.
Click the green Code button, then Download ZIP.
Unzip it somewhere on your machine. This folder is your project root for the rest of this appendix and for every notebook in the course.

If you’re already comfortable with git, cloning works the same way and makes it easier to pull future updates to the course material:

git clone https://github.com/MatthBogaert/IDAFBE-book.git

A.2 Create a virtual environment

A virtual environment is a self-contained directory that holds a Python installation for a particular version of Python, plus a specific set of additional packages. Without one, packages installed for different projects can silently conflict with each other: a version of pandas that one project needs might break another. Learning to use virtual environments now is a worthwhile habit: it’s exactly how you’ll be expected to manage Python projects later, e.g. in a company.

In this book conda is used to create the environment (you should already have conda available if you installed Anaconda). This book pins everything to Python 3.11: see requirements.txt in the project root for the exact package versions.

A.2.1 VS Code

If you’d like a visual walkthrough of creating a conda environment from within VS Code, this video covers the general steps. Note that the video uses its own example environment name and Python version: when it gets to conda create --name mynewenv, use the command from the next section instead, with this book’s environment name and Python version.

A.2.2 Command line

Open a terminal (not a notebook cell) and run:

conda create --name DA2026 python=3.11

Then activate it:

conda activate DA2026

You may need to restart VS Code (or your notebook) afterward and explicitly select this new environment as your kernel: see Selecting the right kernel below.

A.3 Installing required packages

With the DA2026 environment activated, install the exact package versions this book was written and tested against:

pip install -r requirements.txt

Make sure requirements.txt is the one in the root of this book’s project folder: running this from the wrong directory, or skipping it entirely and just pip install-ing packages one at a time, is the single most common source of “it works in the book but not on my machine.”

Installing the right versions

You’ll sometimes be tempted to just install the newest version of a package. We deliberately don’t: for example, pandas 3.0 was released recently, but this book stays on pandas>=2.3,<3 because several packages it depends on (feature_engine, sklearn_evaluation) haven’t been fully verified against pandas 3.0’s changed default behavior yet. See the note in Python for Data Analytics for details. The pinned versions in requirements.txt are the ones every example in this book was actually run against.

If you’re running this from inside a Jupyter notebook rather than a terminal, you can install the same requirements with:

%pip install -r requirements.txt

A.4 Selecting the right kernel

Installing the packages into the DA2026 environment isn’t enough on its own: VS Code (or Jupyter) also needs to actually use that environment when it runs your code. This trips people up more often than the installation itself, so it’s worth being deliberate about it.

For .ipynb notebooks: open the notebook, click the kernel name in the top-right corner of the notebook editor, choose “Select Another Kernel…”, and pick “Python (DA2026)” from the list.

For your overall Python interpreter (affects linting, IntelliSense, and some extensions): Cmd+Shift+P (or Ctrl+Shift+P on Windows/Linux) → “Python: Select Interpreter” → choose the interpreter inside your DA2026 environment.

For quarto preview/quarto render on .qmd files: this project’s _quarto.yml already pins jupyter: da2026 at the top, so Quarto picks the right kernel automatically regardless of whatever interpreter VS Code itself is set to. If a preview still seems to be using the wrong environment, close and reopen the preview tab: it can cache the kernel it started with from before you had DA2026 set up.

The conda environment trap

If you have multiple conda environments with the same Python version, they can look identical in VS Code’s picker unless you check the environment name specifically, not just the version number. Picking the wrong one is easy to do, and the symptom is usually confusing: code that runs fine in one environment throws an error like AttributeError: _ARRAY_API not found or A module that was compiled using NumPy 1.x cannot be run in NumPy 2.x in another. That specific error means the active kernel has a numpy version that doesn’t match the pandas/pyarrow build it’s trying to import: you’re not actually in DA2026. Re-check the kernel/interpreter selection above.

A.5 Running Python

Once your kernel is set correctly, you’re ready to run code. As a first check, try the classic “Hello, world!” example: Python’s print() function displays text or results.

print("Hello, world!")

Hello, world!

If that runs without error, your environment is set up correctly and you’re ready for the rest of this book.