I have created a visualization based on the U.S Dept of Education College Scorecard database that, among other things, allows a user to select a given college or university from a map. That filter action then populates a set of tables that provide information about that university, such as tuition costs, graduation rates, and the percentages of students registered within each major. In the source data, the level of record is a college, and the percentage of students enrolled within each major are separate fields (e.g., "Biology", "Psychology", "Business", etc.). Currently my visualization displays the same six majors, however, what I would like to do is to evaluate the values within those 30+ fields, and to dynamically display only the five largest values. So, for example, at College A, those might correspond to Business, Psychology, English, Physics and Computer Sciences. However, if I selected College B, the five most common majors might be entirely different. Essentially, what I want, is to rank the values of those 30+ fields, then dynamically display only the 5 largest values in ranked order, based on the filter selection of a specific college. I have found plenty of sources for ranking the top N or bottom N values of a single measure based on a selected dimension, but cannot seem to find any way to do this when the measures are all different. Is this possible?
Are you able to reshape your data to bring it all into 1 measure? As you mentioned this will make the Top N quite straightforward.