This is my simple attempt at recreating transfer learning.
As you may see, to make the target course data a bit more realistic, I treated the target course as if they never took quiz 4 like the source course, making the values all 0 as similarly done in the research study when a feature was missing from one of the paired courses.