Thoughts about rand_unique, fct_lvl, lab_lvls, and haystack

[DesiQuintans]

• Move all of these text columns to the end of the dataframe, or at least behind other columns that are shorter and possibly more informative.
• Does haystack really have to exist in the dictionary? Does it save a lot of time to pre-calculate it compared to pasting together all of the other text fields that are already there?
• Factor variables do not need all three of rand_unique, fct_lvl, and lab_lvls filled out.