Setting Up a Python Environment for Data Science

Before analysing data, you need a working environment. This is the setup I followed at the start — install the tools, verify they run, then move on to learning. Steps below cover Windows, macOS, and Linux.

What you are installing

Step 1 — Install Python

Windows

1. Download the installer from python.org/downloads
2. Run it and check “Add python.exe to PATH” at the bottom of the first screen — easy to miss, but important
3. Click Install Now

Open Command Prompt or PowerShell and check:

python --version

You should see something like Python 3.12.x.

On Windows the command is usually python, not python3. If python does not work, try py --version instead.

macOS

Download from python.org or use Homebrew:

brew install python
python3 --version

Linux

Python is often pre-installed. If not:

sudo apt update && sudo apt install python3 python3-venv python3-pip   # Ubuntu / Debian
python3 --version

Step 2 — Create a virtual environment

A virtual environment is a separate folder for packages so they do not clash with other projects.

Windows (Command Prompt or PowerShell):

mkdir my-data-science
cd my-data-science
python -m venv venv
venv\Scripts\activate

macOS / Linux:

mkdir my-data-science
cd my-data-science
python3 -m venv venv
source venv/bin/activate

When active, your terminal shows (venv) at the start.

To leave the environment later, run deactivate.

Step 3 — Upgrade pip and install packages

pip is Python’s package manager. Upgrade it first to avoid common warnings:

python -m pip install --upgrade pip

Use python -m pip instead of typing pip alone — works reliably on every OS.

Then install the core data science stack:

python -m pip install pandas numpy matplotlib seaborn jupyter

Step 4 — Launch Jupyter

jupyter notebook

This opens a browser tab. Click New → Python 3 to create a notebook. Each grey box is a cell — type code, press Shift + Enter to run it.

If jupyter is not recognised, try:

python -m jupyter notebook

Step 5 — Quick verification

Run this in your first cell:

import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt

print("Setup complete")
print("Pandas version:", pd.__version__)

If no errors appear, your environment is ready.

Common issues

python or pip not found (Windows) — reinstall Python and make sure “Add python.exe to PATH” was checked. Close and reopen your terminal after installing.

PowerShell blocks activation (Windows) — if venv\Scripts\activate fails, run this once in PowerShell as admin:

Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser

Or use Command Prompt instead of PowerShell.

pip not found (macOS / Linux) — use python3 -m pip instead of pip.

Permission errors — make sure your virtual environment is activated before installing. You should see (venv) in your prompt.

Yellow pip warnings — usually safe to ignore after upgrading pip. If a package fails, read the last few lines of the error; it often names the missing dependency.

Jupyter won’t open — try python -m jupyter lab as an alternative, or reinstall with python -m pip install --upgrade jupyter.

What comes next

With Python running and libraries installed, you are ready to load a real dataset and explore it. That is where pandas, charts, and groupby come in — the fun part.