Preset Distribution of Sets

I don’t understand that - where does it derive the weighting from?

I am trying to mock a shop branch table. I have a dataset of city names and their population. I want cities with a higher population to appear more frequently in the branch table.