Picks & Shovels

What the competition is, how Kaggle works, and how to register. Bring a laptop: we make Kaggle accounts together and form teams on the spot. The dataset and the competition page go live.

Zero to a submitted baseline. Loading the data with pandas, a first look at the columns, a trivial model, and the exact clicks to submit on Kaggle.

Exploratory analysis: distributions, missing values, what the target actually measures, and what the evaluation metric rewards and punishes.

Turning raw columns into useful features, then fitting the classic baselines: linear models and gradient-boosted trees.

The most important hour of the competition. Cross-validation, leakage, why your public score is not your private score, and how to pick final submissions.

No lecture. Office hours, ensembling tricks, and help choosing which submissions to lock in before the Assay.

Submissions close at the end of the day, Pacific time. The private leaderboard is revealed on Kaggle once the deadline passes.

Awards, prizes, and short talks from the top teams about what worked and what did not.