Beverage Quality Classifier 🥤

A machine learning web application that classifies beverage quality based on various characteristics. Built with Streamlit for the frontend, FastAPI for the backend, and scikit-learn for machine learning.

Features

• 🤖 Machine Learning Classifier: Random Forest model trained on synthetic beverage data
• 🎨 Interactive Frontend: Streamlit web interface with sliders for feature input
• 🔧 REST API: FastAPI backend for serving predictions
• 💾 Database Storage: SQLite database for storing training data and predictions
• 📊 Visualizations: Interactive charts showing feature profiles and prediction probabilities
• 📈 Statistics Dashboard: View database statistics and quality distributions

Beverage Features Analyzed

• Sweetness (0-10): How sweet the beverage tastes
• Acidity (0-10): Level of acidity
• Bitterness (0-10): Level of bitterness
• Carbonation (0-10): Amount of carbonation/fizziness
• Alcohol Content (0-15%): Percentage of alcohol by volume
• Temperature (0-25°C): Serving temperature
• Clarity (0-10): How clear/transparent the beverage is
• Aroma Intensity (0-10): Strength of the beverage's aroma

Quality Classifications

• Poor: Below average quality, needs significant improvement
• Fair: Acceptable quality with room for improvement
• Good: Above average quality, quite enjoyable
• Excellent: Outstanding quality with exceptional characteristics

Installation & Setup

1. Clone the repository:

   git clone <repository-url>
   cd Refreshments-Classifier

2. Run the setup script:

   chmod +x setup.sh
   ./setup.sh

   Or manually:

   pip install -r requirements.txt
   python ml_classifier.py

Usage

1. Start the API Server

python api.py

The API will be available at http://localhost:8000

2. Start the Streamlit App

In a new terminal:

streamlit run streamlit_app.py

The web interface will be available at http://localhost:8501

3. Using the Application

1. Navigate to the Classifier page
2. Adjust the sliders to match your beverage's characteristics
3. Click "Predict Quality" to get the classification
4. View results including:
   • Quality prediction (Poor/Fair/Good/Excellent)
   • Confidence level
   • Probability distribution across all quality levels
   • Feature profile radar chart

API Endpoints

• GET /: API information
• POST /predict: Predict beverage quality
• GET /health: Health check
• GET /feature-info: Get feature information and ranges

Example API Usage

import requests

# Predict quality
features = {
    "sweetness": 7.5,
    "acidity": 3.2,
    "bitterness": 2.1,
    "carbonation": 8.0,
    "alcohol_content": 5.5,
    "temperature": 4.0,
    "clarity": 9.0,
    "aroma_intensity": 8.5
}

response = requests.post("http://localhost:8000/predict", json=features)
prediction = response.json()
print(f"Quality: {prediction['quality']}")
print(f"Confidence: {prediction['confidence']:.2%}")

Project Structure

Refreshments-Classifier/
├── ml_classifier.py      # Machine learning model and data generation
├── api.py               # FastAPI backend server
├── streamlit_app.py     # Streamlit frontend application
├── requirements.txt     # Python dependencies
├── setup.sh            # Setup script
├── README.md           # This file
├── models/             # Saved ML models (created after training)
│   ├── beverage_classifier.pkl
│   └── scaler.pkl
└── beverage_data.db    # SQLite database (created after setup)

Technical Details

Machine Learning Model

• Algorithm: Random Forest Classifier
• Features: 8 numerical features representing beverage characteristics
• Target: 4-class quality classification
• Training Data: 2000 synthetic samples with realistic feature correlations

Data Generation

The synthetic data is generated with realistic correlations between features and quality:

• Higher clarity and aroma intensity correlate with better quality
• Extreme values in acidity or bitterness may reduce quality
• Optimal temperature ranges enhance quality
• Balanced feature combinations result in higher quality ratings

Database Schema

• beverages: Training data with features and quality labels
• predictions: User predictions with timestamps and confidence scores

Future Enhancements

• Add real beverage data collection
• Implement model retraining functionality
• Add data export capabilities
• Include more beverage types and categories
• Add user authentication and personal prediction history
• Implement A/B testing for model improvements

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

License

This project is open source and available under the MIT License.