I was alerted to today’s big news about the iPhone when richf posted about it on webdoc.com. His post lead me to the article on ReadWriteWeb that describes how iPhone software is automatically keeping a log of the phone’s location data and storing it in unencrypted files that are copied to any computer that synchronizes with the iPhone.
At this time, the location file does not seem to be sent to Apple, and the reason for the location tracking remains unknown, but it is very troubling that Apple would implement such a potentially harmful feature without protecting the data and without requiring user consent (or at the very least making a disclosure). I’m very disappointed at this news, because I’m otherwise a happy iPhone user (especially after spending 10 days with a Nexus during a trip to the US last month). Let’s hope it’s simply an very unfortunate error, although an error of such proportions should be more than enough to shake the confidence that most iPhone owners seem to have in Apple.
UPDATE (April, 21, 2:43 GMT) via a new post by richf on webdoc, I’ve read this excellent post by Alex Levinson who has been studying iOS Forensic Analysis for months. Apparently, this is neither a new issue nor a mystery.
Meanwhile, I downloaded the iPhone Tracker application from Pete Warden’s iPhone Tracker site and posted some screensnaps of the resulting location visualizations into richf’s webdoc. I can’t embed that post here, but this is what it looks like:
In comparison with the videos shown on the iPhone Tracker site and in the RWW article, my data seems much more sparse and less accurate. I don’t know yet if this is an artifact of the iPhone tracker app, or if these visualizations accurately reflect the accuracy and time resolution of the recorded data for my phone. Perhaps the triangulation used to determine location does not work as well in Europe as it does in the States, although when I examined data from my recent trip to Austin I did not notice any difference in the quality of the data.
The time granularity of the data seems to vary wildly. I do not know if this is intrinsic to the location tracking implementation or if the tracking requires a data connection. I often turn off International roaming to avoid large data charges when I travel, and this may have had an effect on the sampling. In the past I frequently turned off location services when I wasn’t using them to save battery life, but for the past few months I’ve simply left them on as there are too many apps I use regularly that require location services, and I got tired of always turning them on and off.
UPDATE (April 21, 2:20pm GMT): I found a passage from the iPhone Tracker FAQ that explains this:
To make it less useful for snoops, the spatial and temporal accuracy of the data has been artificially reduced. You can only animate week-by-week even though the data is timed to the second, and if you zoom in you’ll see the points are constrained to a grid, so your exact location is not revealed. The underlying database has no such constraints, unfortunately.
Actually, I remember having read that before installing the app, but I must’ve been a bit too tired to remember it (or find it again) when I was typing this post last night.
This news might provide another reason it’s a good idea to always turn off services you aren’t using, although it’s a shame things have come to that. I’ll be interested to see how Apple responds to this one.
Update (May 2nd): Please see my Detailed analysis of iPhone location data for a more detailed comparison between GPS, iPhone Tracker and consolidated.db location information as well as a test of some of the hypotheses discussed in this post and around the web.