An elevator serving 50 stories should count more than one stopping at only 10 stories. A better measure would be total elevator stops but as usual you are stuck with the available data - especially difficult to overcome with mult-country data like this.

The solution would be to count meters of elevator per person.

That's still not good enough because a tiny one person elevator shouldn't count the same as a massive 30 person elevator.

Hilarious! I have had similar discussions - just about different data source.

Or perhaps person-distance/time or mass-distance/time of capacity?

Mass-distance, sure. Time starts becoming a metric of the quality and vintage of the elevators involved, and not of the capacity.

I don't remember a place where a single elevator serves 50 floors

I've seen buildings where each elevator (set) covered around 12 floors (+ ground floor). This was for a building with +/- 40 floors

That's mostly for efficiency. Imagine going on the top floor and having to wait for 10+ stops on the way up. That said, I think split elevators are more common in commercial buildings. I've seen residential skyscrapers have elevators cover all floors. For example, One Rincon Hill in San Francisco has 3 elevators that all cover all 64 floors.

The five elevators in my building (Shenzhen) serve all 45 floors and it's a mess. I think I must have spent 10+ minutes waiting for the elevator once.