Substitute the following to make an exploratory data analysis version:

  1. Data is still (always) a giant block of mess.
  2. Data is often cleaned instead of sorted. Data cleaning includes things like recoding and removing invalid values.
  3. Data is arranged by feeding it through some fancy mathy code that looks for patterns in it.
  4. Making the viz is the easy part, finding the right viz though can be terribly difficult.

Source: Hot Butter Studio