Conundrum 29: Distilling Data

[CoreyS]

Welcome to Conundrum 29!

Today we take a look at a dataset of global universities ranked by various metrics within their respective nations.

In order to predict the quality of each university in a future week we will first need to distil down a single metric for each, generically termed ‘quality’.

Can you use the various datapoints given to create an overall ‘quality’ score for each university? Note that most of the metrics available are global rankings, from four separate ranking institutions (so expect four rank 1s etc). 

Good luck!

[Ignacio_Toledo]

Hi @CoreyS! This looks like a nice dataset to work with!

I've downloaded the dataset and I wonder if you have more information about the values of the different columns: are they ranks? or totals? or some kind of unknown metric?

Also, some institutions are repeated up to 4 times, but some of the columns values are different: are those duplicates entries from other years?

Thanks in advance!

I.

[MichaelG]

Hi @Ignacio_Toledo - thanks for the question!

They are all international ranks (with the clear exception of 'national rank') - and there should be four first places, four second etc. This is because the data is drawn from four separate sources to give a sense of balance to the data - especially since each source has their own problems of regional and linguistical bias. 

Thanks again for asking, I agree that the original question wasn't to clear (my fault, Corey was kind enough to help out with posting) I'll redraft the original.

[Ignacio_Toledo]

Thanks @MichaelG, that's the info I was looking for.

Cheers!