I'm working in ArcMap 10.1. Currently examining the segment of states with a geographical hierarchy of
1. Counties, coextensive with state boundaries
2. Within counties, unincorporated areas and incorporated places
But incorporated places are in most of these cases not restricted to one county; i.e., a given city can span multiple counties. Although these are to some extent exceptional -- most cities are entirely contained within one county -- not all are of this type.
I know how to do an intersect, calculate area-weighted populations for the resulting segments of places within counties, etc. But I'm wondering if anyone could give me some feedback on how to take an orderly approach to deducting, from the county population numbers, the population of incorporated places lying within, either area-weighted (where the place spans more than one county), or entire: the result I'm looking for would be the population of the balance of the county (i.e., the population of unincorporated areas).
By orderly approach, I mean, how should I manage and reshape the data, use database operations, etc. The process of doing what I'm talking about is pretty straightforward when I'm looking at a small set of features and can analyze them piece-by-piece, but I'm looking at a huge set of features.
It may be worth noting that the attribute tables for the places layers do not have a foreign key field for the counties in which they lie, which, perhaps, was obvious from the fact that they may span multiple counties.
My apologies if I've set this forth in a confusing way. Thanks in advance for any help you could provide!