Hm. While I didn't look directly at what was in the database at that point, I modified Peter Warden's tool to let me get a more refined picture of what was going on, both by increasing the precision, and by aggregating the data on a daily basis, as opposed to a weekly one.

That said, I actually got into what gets stored in a representative consolidated.db's "CellLocations" table today, and dumped out raw data to an Excel spreadsheet (which you can also download if you want to play with it yourself). I also mapped each "batch" of locations—since they're time-stamped in a dozen batches ranging from 7 to 43 locations, collected at intervals ranging from 1 minute to over 16 hours, and covering geographical areas from a few blocks up to 80 or 90 miles. You can see all the details at http://caffeine.shugendo.org/2011/04/22/an-even-deeper-dive-into-the-iphone-location-data/

Now that we've gotten that out of the way, kiss my ass.