Data munging is the process of cleaning or formatting data so that it will serve a model, graphic or any other use. It often refers to making changes to the dataset whenever problems are found. A few examples of processes that could be considered data munging: dropping empty variables (columns), formatting text or numbers, sorting.