LIME Wine

I like wine. I like wine data. I don't understand wine reviews.

What makes a wine good or bad, according to the professionals? We’re going to take 130 thousand wine reviews and train a fancy model to find out. We’ll use Locally Interpretable Model-agnostic Explanations (LIME) to describe how the model treats individual reviews, complete with a Shiny interface. Is it a good thing for wine to have the “aromas of tennis ball and chalk”? Should you bring that bottle to a dinner party if it tastes like “a pot of stewed prunes”? Let’s find out!

Dr. David Neuzerling is a Melbourne-based data scientist. After finishing his doctorate in abstract algebra he jumped ship to industry. He’s used data to tell stories in finance, utilities and sales, and likes to make machine learning understandable to humans. Apart from being a data nerd, David likes dogs, fountain pens, and improvisational comedy.

Pizzas and beers will be provided.