Here is simple way to help organize who is where and help projects like Katrina people finder:

1) News crews video tape people speaking their name, where they are from, and where they currently are (and date). Perhaps additional information.

2) Video sent to closed captioning service to caption (automated/manuel)

3) Google indexes video via closed captioning.

Voila – images & video and information on everyone displaced from storm.

I think we may be doing something closer to indexing than you think. I’m simply assigning chunks of bulletin boards to people. People are then reading those bulletin boards and creating records in a database. That was a standalone SQL database – now it’s something run by Salesforce.com, probably a beast of an Oracle db. So we’re not really aggregating, except to assign people chunks of existing systems to tag and code into a database…

I’m going to suggest that a better approach might be an indexing or searching system, rather than actual aggregation, given the complexity of the free-form data.

That said, I did some work on scraping this sort of stuff last year and will take a shot at one of those sites, if you’ll point me at it. (I suspect it’ll have to be customized for almost every site.)