🍷 Wine Quality Prediction Model

A high-performance machine learning model for predicting wine quality using data streaming for efficient memory management. Combines supervised learning techniques including SVM, XGBoost, and Logistic Regression for robust predictions.

---

🎯 Project Highlights

✨ Memory-Efficient: Streams data using HuggingFace Datasets instead of loading entire dataset into memory 🚀 Multi-Model Ensemble: Combines SVM, XGBoost, and Logistic Regression for superior accuracy 🔬 Supervised Learning: Leverages advanced feature scaling and model optimization techniques 📊 Real-Time Analysis: Visualize predictions with matplotlib and seaborn ⚡ Fast & Scalable: Handles large datasets without consuming excessive storage

---

📊 Quick Stats

Metric	Details
Models Used	SVM, XGBoost, Logistic Regression
Dataset Size	6,400+ samples (via HuggingFace streaming)
Input Features	11 physicochemical properties
Target Variable	Wine Quality (Classification)
Data Source	HuggingFace - mnemoraorg/wine-quality-6k4
Memory Approach	Streaming (Zero-Copy)

---

🔬 Technical Overview

Supervised Learning Approach

This project implements a supervised learning pipeline that:

1. Streams data from HuggingFace Datasets to minimize memory consumption
2. Preprocesses features using MinMaxScaler for normalization
3. Splits data into training and testing sets
4. Trains multiple models (SVM, XGBoost, Logistic Regression)
5. Evaluates performance using comprehensive metrics
6. Visualizes results with matplotlib and seaborn

Architecture

Data Streaming (HuggingFace)
         ↓
Data Exploration & Analysis
         ↓
Feature Scaling (MinMaxScaler)
         ↓
Train/Test Split (80/20)
         ↓
Model Training:
├── SVM (Support Vector Machine)
├── XGBoost Classifier
└── Logistic Regression
         ↓
Model Evaluation & Metrics
         ↓
Visualization & Results

---

📁 Project Structure

wine-quality-prediction/
│
├── wine_quality.py           # Main model implementation
├── README.md                 # This file
└── requirements.txt          # Project dependencies

File Descriptions

wine_quality.py

• Streams wine quality dataset from HuggingFace
• Implements data preprocessing pipeline
• Trains multiple classification models
• Generates performance metrics and visualizations

---

🛠️ Technologies & Libraries

numpy              - Numerical computing
pandas             - Data manipulation and analysis
matplotlib         - Visualization
seaborn            - Statistical data visualization
scikit-learn       - Machine learning algorithms
xgboost            - Gradient boosting classifier
datasets           - HuggingFace Datasets (streaming)

---

🚀 Installation

Prerequisites

• Python 3.8 or higher
• pip package manager
• Internet connection (for dataset streaming)

Step 1: Clone the Repository

git clone https://github.com/mwirigijustice101-pythonist/codespaces-blank.git
cd codespaces-blank

Step 2: Create Virtual Environment

# Using venv
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Or using conda
conda create -n wine-quality python=3.9
conda activate wine-quality

Step 3: Install Dependencies

pip install -r requirements.txt

---

📖 Usage

Quick Start

# Run the wine quality prediction model
python wine_quality.py

What the Script Does

1. Streams Dataset from HuggingFace

   dataset = load_dataset("mnemoraorg/wine-quality-6k4", split="train", streaming=True)

2. Previews Data (first 3 examples)

   for row in dataset.take(3):
       print(row)

3. Converts to Pandas DataFrame

   df = dataset.to_pandas()
   print(df.head())

4. Preprocesses Features

   scaler = MinMaxScaler()
   X_scaled = scaler.fit_transform(X)

5. Trains Multiple Models

   svm_model = SVC()
   xgb_model = XGBClassifier()
   lr_model = LogisticRegression()

6. Evaluates & Compares Results

   from sklearn import metrics
   accuracy = metrics.accuracy_score(y_test, predictions)

---

💡 Key Features

1. Data Streaming 🌊

Loads data efficiently from HuggingFace without consuming local storage:

from datasets import load_dataset
dataset = load_dataset("mnemoraorg/wine-quality-6k4", split="train", streaming=True)

Benefits:

• ✅ Memory efficient
• ✅ Scalable to large datasets
• ✅ No local storage required
• ✅ Real-time data access

2. Multiple Classification Models 🤖

Support Vector Machine (SVM)

• Non-linear classification
• Optimal for high-dimensional data
• Robust against outliers

XGBoost Classifier

• Gradient boosting implementation
• Fast training and inference
• Handles non-linear relationships

Logistic Regression

• Baseline model for comparison
• Interpretable results
• Fast training

3. Feature Scaling 📊

Normalizes all features to [0, 1] range:

from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
X_scaled = scaler.fit_transform(X)

4. Comprehensive Evaluation 📈

• Accuracy score
• Precision, Recall, F1-Score
• Confusion matrix
• ROC-AUC score
• Classification reports

5. Data Visualization 📉

import matplotlib.pyplot as plt
import seaborn as sb

# Visualize model performance
plt.figure(figsize=(10, 6))
# ... plotting code ...
plt.show()

---

🔍 Model Comparison

Model	Strengths	Use Case
SVM	Excellent for binary/multiclass, handles non-linear patterns	When you need robust, accurate boundaries
XGBoost	Fast, handles imbalanced data, feature importance	Production-grade predictions
Logistic Regression	Interpretable, baseline model, probability outputs	When model explainability matters

---

📊 Dataset Information

Source

🔗 HuggingFace Datasets: mnemoraorg/wine-quality-6k4

Features (11 Physicochemical Properties)

1. Fixed Acidity - tartaric acid concentration (g/dm³)
2. Volatile Acidity - acetic acid concentration (g/dm³)
3. Citric Acid - citric acid concentration (g/dm³)
4. Residual Sugar - sugar remaining after fermentation (g/dm³)
5. Chlorides - sodium chloride concentration (g/dm³)
6. Free Sulfur Dioxide - molecular SO₂ (mg/dm³)
7. Total Sulfur Dioxide - bound and free SO₂ (mg/dm³)
8. Density - wine density (g/cm³)
9. pH - acidity level
10. Sulphates - potassium sulphate concentration (g/dm³)
11. Alcohol - alcohol content (% vol)

Target

Quality Score - Discrete values from 0-10 (classification problem)

---

📦 Requirements

Create a requirements.txt file with:

numpy>=1.21.0
pandas>=1.3.0
matplotlib>=3.5.0
seaborn>=0.11.0
scikit-learn>=1.0.0
xgboost>=1.5.0
datasets>=2.0.0
scipy>=1.7.0

Install All Dependencies

pip install -r requirements.txt

---

🎓 How to Extend This Project

1. Add More Models

from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier

rf_model = RandomForestClassifier(n_estimators=100)
gb_model = GradientBoostingClassifier()

2. Implement Hyperparameter Tuning

from sklearn.model_selection import GridSearchCV

param_grid = {'C': [0.1, 1, 10], 'kernel': ['linear', 'rbf']}
grid_search = GridSearchCV(SVC(), param_grid, cv=5)

3. Add Cross-Validation

from sklearn.model_selection import cross_val_score

scores = cross_val_score(model, X, y, cv=5)
print(f"Cross-validation scores: {scores}")

4. Create Prediction Function

def predict_wine_quality(features):
    """
    Predict wine quality from features
    Args: features (array-like) - 11 physicochemical properties
    Returns: prediction (int) - predicted quality score
    """
    X_scaled = scaler.transform([features])
    return model.predict(X_scaled)[0]

5. Save Trained Models

import pickle

# Save model
with open('wine_model.pkl', 'wb') as f:
    pickle.dump(model, f)

# Load model
with open('wine_model.pkl', 'rb') as f:
    loaded_model = pickle.load(f)

---

🎯 Best Practices Implemented

✅ Supervised Learning: Properly labeled training data ✅ Data Streaming: Memory-efficient HuggingFace Datasets integration ✅ Feature Scaling: MinMaxScaler for normalized features ✅ Train/Test Split: 80/20 split for unbiased evaluation ✅ Multiple Models: Ensemble approach for robustness ✅ Warnings Suppressed: Clean console output ✅ Modular Code: Easy to extend and maintain

---

📈 Performance Metrics

The script evaluates models using:

• Accuracy: Overall correctness
• Precision: True positive rate among predicted positives
• Recall: True positive rate among actual positives
• F1-Score: Harmonic mean of precision and recall
• Confusion Matrix: Detailed classification breakdown
• ROC-AUC Score: Area under receiver operating characteristic curve

---

🐛 Troubleshooting

Issue: Dataset Download Takes Long

Solution:

• First run caches the dataset locally
• Subsequent runs load from cache
• Ensure stable internet connection

Issue: Memory Issues

Solution:

• Data streaming already minimizes memory usage
• If issues persist, process data in smaller batches:

  batch_size = 1000
  for batch in dataset.batch(batch_size):
      # Process batch

Issue: Import Errors

Solution:

• Verify all dependencies installed: pip install -r requirements.txt
• Check Python version: python --version (should be 3.8+)

Issue: Model Training is Slow

Solution:

• Reduce dataset size for initial testing
• Use n_jobs=-1 for parallel processing:

  SVC(kernel='rbf', n_jobs=-1)

---

🤝 Contributing

Contributions are welcome! Here's how:

1. Fork the repository
2. Create a feature branch: git checkout -b feature/my-feature
3. Make changes and test thoroughly
4. Commit: git commit -m "Add my feature"
5. Push: git push origin feature/my-feature
6. Open a Pull Request

---

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

🙏 Acknowledgments

• Dataset Source: HuggingFace Datasets - mnemoraorg/wine-quality-6k4
• Original Data: UCI Machine Learning Repository
• Libraries: scikit-learn, XGBoost, pandas, numpy communities
• Research: [Cortez et al., 2009] - Wine Quality Dataset

---

📚 References

1. Cortez, P., Cerdeira, A., Alves, F., Matos, T., & Reis, J. (2009). Modeling wine preferences by data mining from physicochemical properties. Decision Support Systems, 47(4), 547-553.

2. HuggingFace Datasets Documentation

3. scikit-learn Supervised Learning

4. XGBoost Documentation

---

📞 Support & Contact

• GitHub Issues: Report issues here
• Author: @mwirigijustice101-pythonist

---

🚀 Ready to predict wine quality? Clone and run!

git clone https://github.com/mwirigijustice101-pythonist/codespaces-blank.git
cd codespaces-blank
pip install -r requirements.txt
python wine_quality.py

⭐ If you found this helpful, please star the repository! ⭐

Made with ❤️ by mwirigijustice101-pythonist

Last Updated: May 11, 2026