I’ve decided to start off the year by publishing a series of videos on getting ahead of the Titanic machine learning competition with, well, with as little effort as possible.

The plan is to have this split into several videos, showing how you can use scikit-learn to train a model, making the switch to Automated ML, understanding what’s going on under the hood, how the generated pipeline works and how you can use it, in whole or just the parts that you like. It should be fun 🤓.

Part 1 - scikit-learn

The first video of the series is below. In it I discuss the Titanic competition, and show one of the simplest (and possibly dumbest) way to train a model for the competition. You can follow along the video using the notebook from github.

Part 2 - Automated ML

Second episode is here - here I show you how to get from top 96% to top 16%, using a basic Automated ML setup. As before, you can follow along the video using the notebook from github.