Its a misleading ranking because Nature journals arent considered top tier publications in many fields, hence it is very biased towards the experimental sciences (and the life sciences in particular) which tend to put Nature above everything, and consider Nature field journals (eg Nature Genetics) to be top tier publications.
If you click on the number to the right of each university, you can see the actual papers that were counted. Around 60-70% of them are life sciences, and another 20% or so come from other experimental sciences (mostly physic). Other fields (humanities, social sciences, mathematics, computer science, theoretical science, etc) get almost no representation.