Jared Wilber posted this great walkthrough where he codes a simple R data pipeline using purrr and tidyr to train a large variety of models and methods on the same base data, all in a non-repetitive, reproducible, clean, and thus tidy fashion. Really impressive workflow!