Hacker News Reader

Comment by petewarden | Hacker News Reader

I'll be checking in here for technical questions. The github direct link is http://petewarden.github.com/iPhoneTracker/

Thanks for the app -- quite eye-opening. The data looks good, except there is a bunch of data in Minnesota. I've never been to Minnesota! Any ideas?

http://imgur.com/ORgYu

EDIT: Data appears around the time I flew from NJ to Seattle -- perhaps my phone was on in-flight and connecting to MN cell towers? Seems unlikely, though, given the # of points.

Also unlikely because if your phone had not been in the "off position" then your plane would have undoubtedly crashed.

That could be related to the iOS location bugs. I know there were a couple where the phone would refuse to believe that it was anywhere but 1000 miles away.

Pete, the iPhone Tracker application is showing data points in a grid-like pattern, with multiple points visible simultaneously.

I haven't dug into the raw data - is that an artifact of the iPhone Tracker application, or is the raw data already rounded by Apple to fit on a grid, with multiple locations having identical timestamps?

How much I care about this depends a lot on the granularity of the underlying data.

From the FAQ:

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.

Although since it's open source, it shouldn't be too hard to remove this restriction.

I just went through the code and it is indeed pretty trivial to remove both restrictions. If you want to do it, in the iPhoneTrackingAppDelegate implementation file:

* Change the precision variable on line 149 to something bigger than 100 (10000 did it for me) - this will mark points in the map with higher precision:

  const float precision = 10000;

* Change the timeBucket variable on line 180 to change the date precision (the app currently gives you weekly based data, but it can go as far a data point every second). You could also just "cheat" and change the notion of weekInSeconds:

  const float weekInSeconds = (7*24*60*60);

Yo! I realise the likelihood of you seeing this is very low, but is there any chance you would do a brief runthrough of how to implement this code? I'm very interested to increase the precision of this app but don't know enough and can't locate the iphonetrackingappdelegate file you speak of..

I'll check back often! Thanks..!

You need a Mac, git, and XCode

in terminal type: 'git clone https://github.com/petewarden/iPhoneTracker.git

open that new iPhoneTracker folder, and load up the xcodeproj ( iPhoneTracking.xcodeproj )

in there go to iPhoneTrackingAppDelegate.m on the left sidebar

from that look at line 179 for changing the time inbetween registering, and look at line 149 for the locational precision

with those changed, hit the play button in the top left corner and the app should build and run

Thanks for posting the instructions. After doing this, it's easy to see that the tracking data is way off base quite a lot. And interestingly, places I visit a lot but never use my location for it has very few data points for. In fact, there's only one even near my office, which I visit nearly 5 days a week. Every week.

1 data point / second would be cool on flights, but of little practical use on the ground and would seem to be data overkill.

If this is only tracking cell towers as has been reported in a few places, why are there so many more dots than cell towers? Actually the FAQ says it is using triangulation, how accurate is that supposed to be?

Thanks for putting it online. I have a similar project not involving the iPhone logs, but I'm looking to using your visualizations to make sense of the data that I've been collecting.

Reply on news.ycombinator.com