This paper covers five stages of the machine learning process: what to do before building the model, reliably building model, model evaluation, fairly comparing models, and reporting the results.