Emotion Detection – Reproducible ML Pipeline with DVC

This project demonstrates how a notebook-based machine learning experiment can be transformed into a fully reproducible MLOps pipeline using DVC, Git, and a Cookiecutter Data Science (CCDS) structure.

The pipeline classifies text into two emotions:

1. Happiness

2. Sadness

The real goal is not just prediction, but engineering discipline: reproducibility, traceability, modularity, and experiment control.