I have to add my twopennyworth here for public viewing. One of my research difficulties, when making a case for whether two named people are the same or different, is the lack of appropriate information about administrative and ecclesiastic divisions, and even streets in a town or city.
Let me explain: administrative regions would have names and boundaries just like ecclesiastical parishes, and yet they would overlap. This means that it's not just a matter of carving-up some maps to associate names with parts of it. The database then has to link classes of region that overlap. This is very important for English villages as the concept of a "parish of settlement" restricted movements quite severely, and so saying that two references were only 10 miles apart could be misleading.
Note that extended families would often use the same names in their own households, thus diluting the fact that the names were used not too far from each other.
I don't know if I'm taking your concept too far, but other contextual information might include whether there was trade between those places, whether they were on a common trade route, and whether the places were associated with the same industry.
Even for streets (if your GIS would get down to that level of detail) there are similar arguments. Just because two houses were only 2 miles apart doesn't mean that they were of the same social class, or that the housing was provided by the same factories, or that they shared the same local drinking establishments. There's associated context that has to be considered. Town localities were quite territorial (some still are), and English streets -- being non-orthogonal -- would have emphasised that.
These requirements were tackled by my STEMMA work. As well as treating places (of different classes) similarly to people -- rather than just equating names to coordinates -- it allowed their history, documents, and "properties" to be recorded in the same database. For instance, being able to characterise various places as related to the knitting trade (a simple user-defined property) allows quite powerful visualisations of it on a map.
