As I see it, 1:10 is a decent average to get the look right.
However, the reality is more complex.
A small village probably has a lower ratio, they're not likely to pack everybody into 2 buildings. There's more sprawl.
A city is likely to have a mix. Very dense areas in the middle, sparse areas on the outskirts (the inner city is always big apartments, the suburbs are always houses).
It's also a matter of culture and economics. The stereotype of "20 immigrants packed into a single apartment" is true down here in Houston. Sometimes it's dense, because the extended family lives there (plus large family sizes). Sometimes it's because a ton of migrant workers take shifts sleeping in the space. The movie From Hell even demonstrates this, when 5 prostitutes rent a room for the night.
The point is, really big families will crank up the density, and really poor communities will crank up the density, as they live in smaller spaces, or share spaces to reduce the cost.
The 1:10 ratio works well enough for that, as you can assume that's an average. Family homes may have a lower density, apartment buildings, castles, mansions will have a higher density (the help lives in the big mansion as well).
I wouldn't assume the medieval stereotype that everyone has big families and cranks out kids like cattle. You're not running a simulation, so model families like real life, it's close enough, and relateable for your players.
Also assume that for any shop, the owner probably lives there with his family, usually in the space above.
