I put up the pdf of the short paper that goes with it. This is largely a (cliche alert) work in progress, and there are a few different projects that the locator will feed into. Here is the abstract:
This short paper constitutes the first phase of a long-term project focused on probing American urban culture by examining the hyperlinks and text of personal weblogs. It discusses methods of extracting geographic location information from weblogs and ways of indexing weblogs to city units. After a brief introduction to the broader research plan, the paper proposes a process to automatically extract geographic information from different weblogs. From both theoretical and practical perspectives, we explain and justify the rationale of using 3-digit zip codes as units for comparing urban cultures. A distribution of American bloggers registered with Livejournal and Diaryland, two popular weblog hosting services, will be presented to demonstrate the geocoding of the blogosphere, and to compare the distribution of these two hosts in terms of concentrations of populations and demographic profiles. Finally, we discuss how to further improve the indexing methods.