Project 5: Wine Quality Prediction

Predict wine quality based on chemical properties. Can be approached as both regression and classification.

Difficulty

Beginner

Time Estimate

2-3 days

Skills You'll Practice

• Regression/Classification
• Feature Selection
• Handling Imbalanced Data
• Model Tuning

Learning Objectives

By completing this project, you will learn to:

• Work with chemical/analytical data
• Handle imbalanced datasets
• Apply both regression and classification
• Select important features
• Tune model hyperparameters

Dataset

UCI Wine Quality Dataset

• Wine Quality Dataset
• Two datasets: red wine and white wine
• Chemical properties as features
• Quality score (0-10) as target

Features:

• Fixed acidity
• Volatile acidity
• Citric acid
• Residual sugar
• Chlorides
• Free sulfur dioxide
• Total sulfur dioxide
• Density
• pH
• Sulphates
• Alcohol

Target:

• Quality (0-10 scale)

Project Steps

Step 1: Load and Explore Data

• Load red and white wine datasets
• Check data shape and basic statistics
• Analyze quality distribution
• Check for missing values
• Explore feature distributions

Step 2: Data Preprocessing

• Handle missing values (if any)
• Check for outliers
• Feature scaling/normalization
• Create binary classification (good/bad wine)

Step 3: Feature Analysis

• Correlation analysis
• Feature importance
• Visualize relationships
• Identify key features

Step 4: Approach 1 - Regression

• Predict quality score (0-10)
• Train regression models:
  • Linear Regression
  • Ridge Regression
  • Random Forest Regressor
• Evaluate using RMSE, MAE, R²

Step 5: Approach 2 - Classification

• Convert quality to binary (good ≥ 7, bad < 7)
• Handle class imbalance
• Train classification models:
  • Logistic Regression
  • Random Forest
  • SVM
• Evaluate using accuracy, precision, recall, F1

Step 6: Model Comparison

• Compare regression vs classification approaches
• Analyze feature importance
• Select best model
• Final evaluation

Expected Deliverables

1. Jupyter Notebook with complete analysis:

   • EDA with visualizations
   • Both regression and classification approaches
   • Model comparison
   • Results and conclusions

2. Analysis Report:

   • Which chemical properties matter most?
   • Can we predict wine quality accurately?
   • Comparison of approaches

Evaluation Metrics

For Regression:

• RMSE (Root Mean Squared Error)
• MAE (Mean Absolute Error)
• R² Score

For Classification:

• Accuracy
• Precision, Recall, F1-Score
• Confusion Matrix

Key Insights to Explore

• Which chemical properties correlate with quality?
• Is there a difference between red and white wine?
• Can we predict quality from chemical properties alone?
• Which features are most important?

Tips

• Try both regression and classification
• Quality scores are imbalanced (few high scores)
• Feature scaling is important
• Visualize correlations
• Compare red vs white wine models
• Try combining both datasets

Resources

• UCI Wine Quality Dataset
• Wine Quality on Kaggle
• Scikit-learn Regression

Next Steps

After completing this project:

• Try advanced feature engineering
• Experiment with ensemble methods
• Move to Intermediate Projects