The discovery of a wine is of greater moment than the discovery of a constellation. The universe is already full of stars. -- Anonymous
Every wine aficionado has experienced the disappointment of purchasing a $15 bottle of wine, only to uncork and aerate it discover it is undrinkable. Wine experts, known as Sommeliers, can accurately rate a wine's quality from its taste and aroma. But unless you are lucky enough to befriend one, you can't avail yourself of their opinion on a wine before you buy it. Though not a Sommelier, I know what I like in a wine. If I were allowed to taste a wine before buying it, I could choose a good bottle every time.
Suppose that instead of experiencing the wine first-hand, you are only given measurements of a set of physico-chemical properties of each wine. Would you be able to tell good wines apart from bad?
There's a data set for that! In this notebook, you are invited to explore the application of machine learning to wine classification by participating in a deep-dive, end-to-end analysis of the UCI red wine data set. Along the way, you'll become familiar with many of the excellent tools from the scikit-learn library, and pick up a number of practical machine learning tips.