University league tables are built on poor data analysis and arbitrary assumptions, but prospective students and their parents often use them. I can see the attraction of league tables, prospective students and their parents want to compare the different universities they are considering, and league tables are by far the most prominent, and easy to access, way to do that. But as someone who teaches data analysis, and has been an admissions tutor, university league tables make me wince.
A couple of years ago I produced a Google Colab Jupyter notebook that would work with the same data as the top 50 universities in The Guardian‘s 2019 league table, and produce another league table, but with the university of your choice as the ‘top’ university. This was just to show that if you know what you are doing, you can make any university apparently the ‘best’.
But although I like both Jupyter notebooks and Google Colab, to run that notebook you do have to run code. And Jupyter notebook only worked on a list of the top 50 universities – out of the 121 in The Guardian‘s league table. No longer. With the aid of streamlit, I have done a web app that works with a list of all 121, and in over 90% of cases*, produces a league table with whatever university you want top. All at the click of a mouse, and without charge! All the code is under the bonnet of the web app. Unlike The Guardian it won’t even cover the bottom half of the screen with a begging message, despite the much more honest data analysis that my web app represents.
So please pop over and see how high your alma mater, employer, favourite university, …, can rise (spoiler: except for a few cases, all the way to the top). Alternatively, if you actually want to compare universities, try the web app done by a student, Ethan Hinton, who worked with me last year. It is very good.
* The algorithm that I use fails for a handful of universities. I need to work a little harder to find league tables with them at the top. I’ll try and do that when I have the time.
Chance & Necessity 