[lakethun]

Your impressions are incorrect. From a comment above:

   Persons under 18 (2010 Census)
   US 23.7% [2]
   SF 13.5% [3]

[2] http://quickfacts.census.gov/qfd/states/00000.html

[3] http://quickfacts.census.gov/qfd/states/06/06075.html

[egometry]

These two posts are not incompatible. That 13.5% of kids needs to live somewhere. I'd bet that a heatmap would show "kid locuses" in some neighborhoods and "kid voids" in others.

[lakethun]

They are not incompatible only if by "tons" you mean "half as many". I would expect some areas of San Francisco to have more kids than others but overall the number is very low.

[geebee]

SF is a small region in a larger urban area. My guess is that you could easily single out dense, urban sub-regions in LA or Chicago that are as childless as SF. SF used to have more kids, but I'd guess that this used to be true of the dense, urban sections of other cities. The difference is, a kid who moves from urban chicago to suburban chicago is still counted in Chicago's stats.

Meanwhile, some districts in SF might have quite a high percentage of children compared to the average for areas of equal density. Not sure of this, but it would be interesting to see.

[2pasc]

Well... Many families are also in the suburbs (South Bay or East Bay) as the weather tends to be better and you can have houses with gardens for the kids to play around (also the school system is less messed up there). Which does not mean that there are not kids-full neighborhoods in the city.