Skip to content
Judith Solomon
Linear RegressionRegressionscikit-learnPython

Wine Quality — Linear Regression

Predicting a quality score from measurable chemistry.

Role
Learning project
When
August 2026
Tools
Python, pandas, NumPy, scikit-learn, Matplotlib, Jupyter Notebook

Description

An applied linear regression project: predicting the quality rating of a wine from its physicochemical measurements — acidity, sugar, sulphates, alcohol and the rest.

This was the step up from textbook regression to a real dataset, where the features are correlated with each other, the target is ordinal rather than continuous, and the model's confidence deserves more suspicion than its output.

What came out of it

Alcohol content against the quality rating. Quality is recorded in whole numbers, which is why the points lie in bands — and is a large part of why a straight-line model can only go so far here.