Finding the most common maximum of columns

The college dataset contains the undergraduate population percentage of eight different races for over 7,500 colleges. It would be interesting to find the race with the highest undergrad population for each school and then find the distribution of this result for the entire dataset. We would be able to answer a question like, "What percentage of institutions have more White students than any other race?"

In this recipe, we find the race with the highest percentage of the undergraduate population for each school with the .idxmax method and then find the distribution of these maximums.

How to do it…

1. Read in the college dataset and select just those columns with undergraduate race percentage information:

   >>> college = pd.read_csv(
   ...     "data/college.csv", index_col="INSTNM"
   ... )
   >>> college_ugds = college.filter(like="UGDS_")
   >>> college_ugds...