Codes

Code for each day

The executed notebook for each day's empirical work, with every output kept in place. The Day 1 notebook works in eight steps. It loads the data, splits it before fitting anything, sets a baseline that predicts no for everyone, fits a logistic regression on the training set, evaluates on validation, chooses a decision threshold, adds the post-call duration field to show what leakage looks like, then opens the test set once. It reads data/bank-additional-full.csv, so keep the CSV in a data folder beside it. The remaining days are released before each session.