Installing Python and Tooling — Python Foundations for Data Work

This is the moment where your laptop stops being a sad spreadsheet machine and becomes a data scientist's spaceship.

---

Why this matters (and why you should care)

If you're taking a course called 'Python for Data Science, AI & Development', the first real-world barrier is not theory — it's tools. You can know linear regression in your bones, but if your environment is broken you will cry in terminal-font. Installing Python and the right tooling gets you from theory to experiments, reproducible analysis, and shiny notebooks.

In this guide you'll learn practical, battle-tested steps to install Python and set up the tools data people actually use: package managers, virtual environments, Jupyter, an editor, and quick troubleshooting.

---

What we'll cover

• How to get Python on Windows, macOS, and Linux
• Package managers: pip vs conda vs pyenv
• Virtual environments: venv, conda envs, pipenv/poetry (short primer)
• Essential packages and Jupyter
• Editor/IDE recommendations and useful extensions
• Common trouble spots and fixes

---

Step 1 — Install Python (the short routes)

Windows

• Easiest: Download the official installer from python.org and check 'Add Python to PATH' during install.
• Alternative for power users: Install Chocolatey and run:

choco install python

macOS

• Easiest: Use Homebrew:

brew install python

• Or download from python.org if you prefer GUI installers.

Linux (Ubuntu/Debian)

sudo apt update
sudo apt install python3 python3-venv python3-pip

After installing, verify:

python --version
python3 --version
pip --version

If the command is python3 on your system, that’s okay — some systems keep python pointing to Python 2 historically.

---

Step 2 — Choose a package + environment strategy (don’t panic; here's a simple rule)

Rule of thumb: For beginners and most data science workflows, use either conda (Anaconda/Miniconda) or Python + venv + pip. Pick one and be consistent.

Option A: Miniconda/Conda (recommended for data)

• Conda manages Python versions and binary packages cleanly (great for numpy/pandas/scikit-learn where compiled libs matter).
• Install Miniconda (lightweight) from the official site.

Quick commands:

conda create -n ds-env python=3.10
conda activate ds-env
conda install numpy pandas scikit-learn jupyterlab matplotlib seaborn

Option B: Python + venv + pip

• Use builtin venv to isolate projects, pip to install packages from PyPI.

python -m venv .venv
# Activate (Mac/Linux)
source .venv/bin/activate
# Activate (Windows PowerShell)
.\.venv\Scripts\Activate.ps1
pip install --upgrade pip
pip install numpy pandas scikit-learn jupyterlab matplotlib seaborn

Quick note on pyenv

• Use pyenv when you need to manage multiple Python versions system-wide (e.g., 3.8 for an old project and 3.11 for a new one). Combine pyenv with virtual environments.

---

Step 3 — Install Jupyter and your notebook tooling

Jupyter is essential for data work — interactive exploration, plots, and documentation in one place.

Using pip:

pip install jupyterlab
jupyter lab

Using conda:

conda install -c conda-forge jupyterlab
jupyter lab

Tip: JupyterLab is the modern interface; you can also install the classic notebook if you prefer.

---

Step 4 — Pick an editor/IDE and recommended extensions

• VS Code (lightweight, extensible): install the Python extension and the Jupyter extension.
• PyCharm (full-featured IDE): great for larger projects.
• Optional: install Git and set up a GitHub account early.

VS Code quick setup:

1. Install VS Code
2. Install the 'Python' and 'Jupyter' extensions
3. Open your project folder, select the interpreter (bottom-right), and choose your venv or conda env

---

Troubleshooting & common gotchas

• PATH not set: If python or pip is not found, add the install directory to PATH or use python3/pip3 commands.
• DLL errors on Windows with compiled packages: use conda, which provides compatible binaries.
• Wrong interpreter in VS Code: select the correct interpreter manually.
• Virtual environment not activated: your prompt should change; if not, activate it explicitly before installing packages.

Quick fix for permissions when pip fails:

# Prefer user-level installs or virtual envs; avoid sudo with pip
python -m pip install --user package-name

---

Best practices to avoid future pain

• Use virtual environments per project. Do NOT install every package globally.
• Pin dependencies with a requirements.txt or environment.yml.
• Use version control (git) from day one.
• Keep your Python version updated for new projects (but match old projects' versions when needed).

Example: save dependencies

# pip
pip freeze > requirements.txt

# conda
conda env export > environment.yml

---

Quick checklist — go-time

• Install Python or Miniconda
• Create & activate a virtual environment
• Install JupyterLab + core libraries (numpy, pandas, matplotlib, scikit-learn)
• Install VS Code and the Python/Jupyter extensions
• Initialize git in your project folder

---

Final takeaways

• Python tooling is an investment. Spend 30–60 minutes now to set it up right and save hours later.
• Choose conda if you want fewer headaches with compiled libraries; choose venv/pip for minimalism and PyPI access.
• Virtual environments = your future sanity. Use them.

Remember: installing Python is not glamorous, but it is the bootstrap ritual that turns your computer into a lab. Start with a clean environment, document your steps, and you'll be doing reproducible data work faster than you can say 'import pandas as pd'.

Tags: beginner, python, data-science, tooling