Data Science Texts

Diversions

NB: We may earn a commission if you buy something via an affiliate link.

Clashing for Cachet: Three Decades of U.S. News and World Report University Rankings

Tied schools' names are randomly ordered for legibility. For example, if two schools are tied for 3rd, one school name will be placed at rank 3, and the other at rank 4; their paths will both be at 3. Mouseover (tap on mobile) individual schools and paths to make them more distinguishable. If you found this graphic interesting, you might enjoy reading Educated: A Memoir, Tara Westover's highly-acclaimed autobiographical account of first entering a classroom at age 17 to earning her PhD from Cambridge University.

Data Notes

Synopsis

Ranking entities as complicated as universities is guaranteed to be a subjective activity. Despite this, the rankings are taken seriously by many, so I thought it would be interesting to see how the rankings have changed over the years.

Weaknesses of the Visualization

I think there are two major weaknesses to this visualization.
  1. The visualization tells you what is happening in the rankings, but not why it is happening.

    There may not be a terribly satisfying reason to why these rankings are changing. I suspect drastic changes in ranking, such as what happened in 2000 to Cal-Tech, are due to changes in the methodology more than changes in the schools.

  2. The data density is pretty low.

    These race-style graphics are in vogue at the moment, and I wanted to try my hand at making and improving on them. I think this is pretty decent, as it incorporates the animated race but still makes it easy to look at the overall history. However, I think I could get a lot more dimensions into the graphic. I could use glyphs instead of just university names and present more information. I may try this with a visualization that emphasizes characteristics of the schools more than just their USNWR ranks.

Data Sources

Acknowledgments

Thanks to Andrew Reiter for compiling the data. This visualization was created using R and D3. The following packages were particularly helpful:

Notes

1 There seems to be some inconsistency in this data as to which years the first couple rankings pertain. I have used the apparent source data years of 1983 and 1985.↩