The group Nature has updated the Nature Index and published a special issue raking cities and metro areas by their scientific output. There is a vast literature on cities as engines of growth and creativity because of economies of agglomeration, proximity and face-to-face contacts, social and cultural diversity etc. It's interesting to see how scientific outputs are clustered around major urban areas in the world.
The method they used to define urban clusters, however, is not entirely clear. It seems a bit odd to me that they put Columbia, Princeton and Yale Universities in the same 'New York' cluster given the distance and commuting ties between these urban areas. My guess is that based on the same criteria, London-Oxford-Cambride (UK) should also be considered as a single cluster but they are treated separately. The same would happen with Sao Paulo-Campinas-Sao Jose dos Campos (Brazil).
In any case, the report brings plenty food for thought, including this brief analysis showing that the top collaborating cities are located in the same country.
