If the data will be used for machine learning, transformations can ... It’s also important to do your exploratory data analysis (step four) before modeling, to avoid introducing biases in ...