Predicting Wine Quality Score Based on Physicochemical Properties
Determining the quality of wine is tricky. It is often based on subjective tastes of individuals judging them. This project aims to understand the effect of a wine’s properties on its quality score.
The data being used here is from the UC Irvine dataset repository which contains information such as alcohol, sulphates, acidity levels, citric acid among other properties in wine samples collected from red Vinho Verde wine samples, from the north of Portugal.
The goal is to model wine quality based on physicochemical tests. Based on our model, we find that all three predictors (alcohol, sulphates, and citric acid) have positive and significant effects on wine quality, suggesting that increasing these components is associated with higher wine quality scores by as much as 0.8 units.