How to Build and Publish a Python Package on PyPI: My Experience with `my_eda_helper`

Publishing your first Python package on PyPI is an exciting milestone for any developer. It’s a great way to share your work with the world, contribute to the open-source community, and showcase your skills. In this blog, I’ll walk you through the step-by-step process I followed to publish my package, my_eda_helper, which simplifies Exploratory Data Analysis (EDA). I’ll also share tips on handling errors, publishing updates, and best practices. Let’s dive in! 🚀

---

What is my_eda_helper?

my_eda_helper is a Python package designed to streamline the EDA process. It provides helper functions for tasks like:

✅ Missing data analysis

✅ Correlation heatmaps

✅ Hypothesis testing

✅ Interactive visualizations

The goal is to reduce repetitive tasks and allow users to focus on deriving insights rather than writing boilerplate code.

---

Step-by-Step Guide to Publishing a Python Package

1. Plan Your Package

Before writing any code, plan what your package will do. For my_eda_helper, I wanted to automate common EDA tasks and make them reusable.

---

2. Organize Your Project Structure

A well-structured project is crucial for maintainability. Your project should have the following structure:

my_eda_helper/                  <-- Root folder
├── my_eda_helper/              <-- Package folder
│   ├── __init__.py             <-- Makes it a package
│   └── core.py                 <-- Main code
├── setup.py                    <-- Package metadata
├── README.md                   <-- Documentation
├── LICENSE                     <-- License file
└── tests/                      <-- Optional: Tests

📌 Explanation:

• __init__.py: Marks the folder as a Python package. Add from .core import * to expose functions.
• core.py: Contains the main code (e.g., functions for EDA).
• setup.py: Defines package metadata and dependencies.
• README.md: Provides installation and usage instructions.
• LICENSE: Defines the package license.
• tests/: (Optional) Contains unit tests for your package.

---

3. Write setup.py

The setup.py file is crucial for packaging your project. Here’s what mine looks like:

from setuptools import setup, find_packages

setup(
    name="my_eda_helper",
    version="0.1.0",  # Start with 0.1.0 for the first release
    author="Your Name",
    author_email="[e-mail]",
    description="A helper package for Exploratory Data Analysis (EDA)",
    long_description=open("README.md").read(),
    long_description_content_type="text/markdown",
    url="https://github.com/shemanto27/eda-helper-py",
    packages=find_packages(),
    install_requires=[
        "numpy",
        "pandas",
        "seaborn",
        "matplotlib",
        "scipy",
        "statsmodels",
        "scikit-learn",
        "wordcloud",
        "IPython",
    ],
    classifiers=[
        "Programming Language :: Python :: 3",
        "License :: OSI Approved :: MIT License",
        "Operating System :: OS Independent",
    ],
    python_requires=">=3.6",
)

📌 Why?

This file tells PyPI how to install and use your package.

---

4. Write a README.md

Your README.md should include:

• A short description
• Installation instructions
• Usage examples

Here’s a snippet from mine:

# My EDA Helper

A Python package to simplify Exploratory Data Analysis (EDA).

## Installation

bash
pip install my_eda_helper

## Usage

python
import my_eda_helper as eda

Missing data analysis

missing_data = eda.missing_info(df)
print(missing_data)

---

5. Build Your Package

Before uploading to PyPI, build your package:

pip install build
python -m build

This creates a dist/ folder containing .tar.gz and .whl files.

---

6. Upload to PyPI

Step 1: Create a PyPI Account

Sign up at pypi.org.

Step 2: Generate an API Token

• Go to Account Settings > API Tokens
• Create a new token

Step 3: Upload Your Package

pip install twine
python -m twine upload dist/*

• Use __token__ as the username.
• Paste your API token as the password.

✅ Your package is now live on PyPI! 🎉

---

7. Test Your Package

To confirm the installation works, install and import it:

pip install my_eda_helper
python -c "import my_eda_helper; print(dir(my_eda_helper))"

---

Handling Errors and Publishing Updates

Common Errors

❌ Package Not Found:

• Ensure the package name is unique on PyPI.
• Double-check the name field in setup.py.

❌ Missing Files:

• Ensure all necessary files (__init__.py, core.py, README.md) are included.

❌ Version Conflict:

• Increment the version number in setup.py before uploading a new version.

---

Publishing a New Version

If you want to update your package:

1️⃣ Update the version number in setup.py (e.g., 0.1.1).

2️⃣ Rebuild the package:

python -m build

3️⃣ Upload the new version:

python -m twine upload dist/*

✅ Done! Your new version is live.

---

Final Thoughts

Publishing a Python package is a rewarding experience. Whether it’s a small utility or a full-fledged library, your work can help others and grow your skills.

If you found this guide helpful, follow me for more tips and projects:

🔗 LinkedIn: https://www.linkedin.com/in/shemanto/

🐍 GitHub: http://github.com/shemanto27/

Happy coding! 🚀