🚀 MLFlow Metrics Tracking Tutorial - MLOps Python Project

A practical tutorial demonstrating MLFlow for metrics tracking and experiment management using a bike sharing dataset.

📋 Table of Contents

• Overview
• Project Structure
• Quick Start
• MLFlow UI
• Tutorial
• Contributing
• Acknowledgments

🎯 Overview

A hands-on tutorial for learning MLFlow metrics tracking and experiment management.

This project demonstrates how to use MLFlow for tracking machine learning experiments using a bike sharing demand prediction dataset. It emphasizes:

• 📊 Experiment Tracking: Learn to track metrics, parameters, and artifacts
• 🔄 Model Versioning: Understand model registry and versioning concepts
• 📈 Metrics Visualization: Compare experiments and visualize results
• 🛠️ Best Practices: Apply MLOps principles in practice

✨ Features

• MLFlow 3.1 Integration: Latest MLFlow features for experiment tracking and model registry
• Enhanced Model Registry: Improved model versioning and lifecycle management
• Real Dataset: Uses bike sharing demand dataset for practical learning
• Jupyter Notebooks: Interactive tutorials and examples
• Model Training: Example ML models with proper logging
• Metrics Tracking: Comprehensive metrics and parameter logging
• Artifact Management: Model and data artifact storage with enhanced metadata

📂 Project Structure

mlflow-1-metrics-tracking/
├── data/                   # Data files
├── dev/                    # Development files
│   └── images/            # Development images
├── docs/                   # Documentation and images
├── mlartifacts/           # MLFlow artifacts storage
├── mlruns/                # MLFlow runs metadata
├── models/                # Trained models storage
├── notebooks/             # Jupyter notebooks for tutorials
│   └── mlruns/           # Notebook-specific MLFlow runs
├── src/                   # Source code
├── README.md              # This file
├── requirements-dev.txt   # Development dependencies
└── requirements.txt       # Project dependencies

🚀 Quick Start

1. Clone the Repository

git clone [YOUR_REPO_URL]
cd mlflow-1-metrics-tracking

2. Set Up the Environment

make setup                  # Create virtual environment
make install-all            # Install dependencies

3. Download the Dataset

# Load bike sharing dataset
python src/load_data.py

4. Verify Installation

# Check if MLFlow is properly installed
mlflow --version

📺 MLFlow UI

Start the MLFlow tracking server:

mlflow server --host 0.0.0.0 --port 5001

Then navigate to http://localhost:5001 in your browser to:

• 📊 View Experiments: Compare different model runs with MLflow 3.1's enhanced UI
• 📈 Analyze Metrics: Visualize training metrics over time with improved charts
• 🔍 Inspect Artifacts: Download models and other artifacts with better metadata
• 📋 Compare Runs: Side-by-side comparison of experiments with advanced filtering
• 🏷️ Model Registry: Manage model versions with enhanced lifecycle stages

🎓 Tutorial

Launch Jupyter Lab to access the interactive tutorials in notebooks/ directory:

jupyter lab

The notebooks will guide you through:

1. Data Exploration: Understanding the bike sharing dataset
2. MLFlow 3.1 Setup: Setting up tracking and logging with latest features
3. Experiment Tracking: Logging parameters, metrics, and artifacts with enhanced metadata
4. Model Comparison: Comparing different models and hyperparameters using improved UI
5. Model Registry: Managing model versions and stages with MLflow 3.1's enhanced lifecycle management
6. Advanced Features: Exploring MLflow 3.1's new capabilities and improvements

🤝 Contributing

1. Fork the repository
2. Create a feature branch: git checkout -b feature/improvement
3. Make your changes
4. Test your changes: Ensure notebooks run correctly
5. Commit your changes: git commit -m 'Add improvement'
6. Push to branch: git push origin feature/improvement
7. Open a Pull Request

Development Guidelines

• Follow Python best practices and PEP 8
• Add type hints to functions
• Document new features in notebooks
• Test MLFlow logging functionality
• Update README if adding new features

🙏 Acknowledgments

• Dataset: The bike sharing dataset is from Kaggle Bike Sharing Demand
• Research: Fanaee-T, Hadi, and Gama, Joao, 'Event labeling combining ensemble detectors and background knowledge', Progress in Artificial Intelligence (2013): pp. 1-15, Springer Berlin Heidelberg
• Data Source: UCI Machine Learning Repository - Bike Sharing Dataset
• Integration Example: Based on the mlflow_monitoring integration example from Evidently AI

---

Happy learning with MLFlow! 🎉