Skip to content

Wine

Wine is the fermented juice of grapes.

Wine can be classified in a variety of ways including color, residual sugar content, carbon dioxide content, alcohol content or graph variety. Generally there are five basic characteristics of wine

  • Sweetness - that is the level of residual sugar left in the wine after its creation
  • Acidity - it gives it sharpness - high acidity wine are often tart and zesty. A well-balanced wine is so called as it has acidity, sweetness and tannin in perfect harmony
  • Tannin - tannin is the presence of phenolic compounds that add bitterness to a wine, and also adds balance ans structure, and helps wine last longer
  • Alcohol: wine alcohol percentage level have the biggest impact on a wine's character, body and classification.
  • Body - it can help to think of a wine's body like milk, with skimmed milk representing a light wine, and cream representing a full-bodied wine. it is said if a wine's taste lingers in your mouth for more than 30 seconds, it's almost certainly a full-bodied wine

As can be seen, these characteristics are very subjective to human nature, and the outcome may differ from one individual to another. Machine learning provides an impartial way to predict the quality of the wine.

We'll work with three types of datasets to solve three different problems.

  1. Wine source classification
  2. Wine Quality prediction
  3. Wine Recommendation

Business Objectives

  1. Enhanced Product Quality: Accurate wine quality prediction will lead to improved quality and consistency, enhancing the winery reputation and customer satisfaction
  2. Cost Optimization: Optimal resources allocation and reduced wastage through predictive modeling will result in cost saving of wineries, improving overall operational efficiency
  3. Market Competitiveness: Consistent production of high-quality wines will give wineries competitive advantage, allowing them to stand out int he market and attract more customers.
  4. Better Customer Experience
  5. Improved Revenue

Project Objective

  1. Develop a predictive model using machine learning algorithms to accurately assess the quality of red and white wines based on various chemical properties and attributes
  2. Evaluate and compare the performance of different machine learning algorithms to determine the most effective approach for wine quality prediction providing insight for potential industry application
  3. Develop an API to consume the model
  4. Package the model for CI/CD

Code: joewynn/mlops-for-fun