Managing a Python Environment for Data Analysis

Learn Pixi by Managing a Python Environment for Data Analysis

Goal

Learn how to use Pixi to create, manage, and reproduce a Python environment using only Pixi’s core features.

What You Will Learn

• How to install and verify Pixi
• How to create a new environment
• How to install packages from conda-forge and PyPI
• How to activate the environment and run commands
• How to export and share the environment

Step 1: Install Pixi

Visit pixi.sh and follow the instructions for your operating system.

Example for Windows:

winget install prefix-dev.pixi

Verify installation:

pixi --version

Step 2: Create a New Pixi Environment

Create a new project directory:

mkdir my-data-env
cd my-data-env
pixi init

This creates a pixi.toml file where your environment is defined.

Step 3: Add Python and pip

pixi add python pip

Verify Python is available:

pixi shell
python --version
pip --version

Type exit to leave the shell.

Step 4: Add Data Packages

Install useful packages:

pixi add pandas matplotlib

Install a PyPI-only package (such as tabulate):

pixi add --pypi tabulate

View environment info:

pixi info

Step 5: Test the Environment

Create a file test.py:

import pandas as pd
from tabulate import tabulate

df = pd.DataFrame({'A': [1, 2], 'B': [3, 4]})
print(tabulate(df, headers='keys', tablefmt='pretty'))

Run it inside the environment:

pixi run python test.py

Step 6: Share or Reproduce the Environment

To share: - Send pixi.toml and pixi.lock

To reproduce on another machine:

pixi install
pixi run python test.py

Summary

You have now:

• Created a Pixi-managed environment
• Installed packages from both conda and PyPI
• Used pixi run and pixi shell
• Reproduced the environment with lock files