Everyblock launched yesterday as a site that shows you news and other item that are about or relevant to your neighborhood. You enter your address, postal code or neighborhood name and see news articles, civic documents like crime reports and building permits, blog posts, craigslist entries, Yelp reviews, and Flickr photos associated with the area around it.
Currently only three cities are covered by everyblock — New York, Chicago and San Francisco. For an example, see the everyblock page for NYC’s Chelsea neighborhood or around the University of Chicago.
I suspect that much of the work that goes into a system like everyblock is selecting the right sources. For example, how you access online local government documents for each state or city will differ. You would want to mine the local newspapers for news items and focusing on them would make disambiguating geonames and addresses easy, since your first order approximation would be that every geo-reference is local. There is also work in adapting to the APIs for other services, like Yelp and craigslist, that each has its own way to sorting items into geographic regions.
There are good data sources for the GIS information and services, free or paid, that will do the geocoding and reverse geocoding of names and addresses. The Earth is a finite place and we have a lot of data about what is where, at least in the more developed parts of it.
Although everyblock claims to include relevant blog posts, I’ve not seen any yet. This is a harder problem, unless you get bloggers to add explicit geographic metadata or register their blogs with a location, like feedmap.org does.