Project to predict which request for free pizza (on reddit) results in a positive (or negative) response

Notebooks with code in main directory where data is ingested, transformed, features are selected (and produced), and finally models are run. We find that simple models do well i.e. logistic regression compared to decision trees, SVM, ensemble models, or linear regression.

Data is in /data directory