Machine Learning in Production with Python

Background Context:

In this code-along session, we will work with a dataset from a Portuguese banking institution's direct marketing campaigns, sourced from Kaggle. The dataset contains information about various marketing efforts, including telephonic outreach, aimed at promoting term deposits. Term deposits are significant for banks as they provide a stable income stream. Identifying and targeting potential customers effectively can enhance marketing efficiency and reduce costs.

https://www.kaggle.com/datasets/prakharrathi25/banking-dataset-marketing-targets/data

Tasks Covered:

1. Feature Engineering:

   • We will develope functions to enhance the DataFrame with new features.
   • These functions were combined into a streamlined data transformation process using the Pandas pipe method.

2. Unit Testing:

   • We implemented unit tests using the pytest framework to ensure the new features were correctly added to the DataFrame.
   • The tests verified that the new columns were accurately calculated and correctly incorporated into the DataFrame.

Overall, this session provided practical experience in enhancing a dataset with valuable features and ensuring the robustness of these enhancements through thorough testing.

To ensure a clean and isolated development environment, use a virtual environment created with venv and manage dependencies using pyproject.toml.

Additionally, set up pre-commit hooks for code quality checks, including linting with Ruff.

Create and activate a virtual environment:

python -m venv venv
venv\Scripts\activate

Once inside the virtual environment, install project dependencies from the pyproject.toml file:

pip install -e .

Poetry and pip-tools are excellent package managers. As pip won't necessarily handle all dependencies and constraints as effectively as package managers

Use pre-commit to enforce code quality standards. Additionally, use ruff for linting. Both packages are in the pyproject.toml file for install.

https://pre-commit.com

Navigate to your project directory and set up pre-commit hooks:

pre-commit install

A .pre-commit-config.yaml file is in project directory with config for ruff linting. Whenever you make a commit, pre-commit will run linting with Ruff and enforce code quality standards automatically.

https://github.com/astral-sh/ruff-pre-commit