3 ms·

Interesting. For me, the data shows a lot of where I've been, but with only the sort of accuracy (and errors) you'd expect from tower triangulation. It also m

by brk 16y ago

Interesting. For me, the data shows a lot of where I've been, but with only the sort of accuracy (and errors) you'd expect from tower triangulation. It also misses some places that I frequent entirely (eg: it has almost no hits near my weekend house or some of the local places there my wife and I visit). Similarly shows no hits in the Pacific Northwest for me, but did seem to track my Vegas and Santa Barbara trips pretty well (though again with highly scattered accuracy). Maybe it's just me, but I'm having trouble feeling very "violated" by this. If you were to look at my map, I think you could tell the area of the country where I live/work, but probably couldn't pick out my house or office with much accuracy (unless there are underlying timestamps the map doesn't show).

willvarfar 16y ago
Yes all the points have timestamps and an accuracy estimate, and there is wifi fix data too that this script doesn't show. There is often erronous data in there too - that's not this script, that's just that location fixes are inexact. The real violation is that you just drag-dropped your whole iphone backup to a script written by people you don't know and running on servers you don't control... but don't worry, we don't care to profile you or sell you data... ;)
- brk 16y ago
  The real violation is that you just drag-dropped your whole iphone backup to a script Yeah, I guess. I just did a quick strings on the backup files, and you could extract a fair bit of info, but it didn't seem like there was much in there that was truly juicy.
schleyfox 16y ago
I'm growing increasingly convinced that the first iphone tracker desktop app reduced precision not so much to "protect privacy", but to cover up exactly how off the data was. I think selling it as though it were your location rather than tower location backs this up. Looking at it in Google Earth it seems to only be accurate when I was on I-95 in New Jersey. I figured the data would be kind of rough, but for a month of my data in NYC it is just off. Apparently I spent most of my time in Bed Stuy and no time in my apartment in Park Slope. It also got Charlottesville, VA confused with Las Vegas. If that actually happened, there must be some great party stories that I haven't been told about yet.
- markolson 16y ago
  Interestingly, that was my first thought too. I'm the guy who hacked together the javascript which passably reads the database file in-browser for this, and two things stood out to me as I was working: How much data there was, and how not-quite-that-accurate it was. I think the initial app's way of presenting the data has merit because it does hide the swarm of identical points, but it also does make it 'feel' more accurate when you look at it because 'Yes! I live in that circle.' There's sure to be a happy middle ground somewhere.

by brk 16y ago

willvarfar 16y ago

Yes all the points have timestamps and an accuracy estimate, and there is wifi fix data too that this script doesn't show. There is often erronous data in there too - that's not this script, that's just that location fixes are inexact. The real violation is that you just drag-dropped your whole iphone backup to a script written by people you don't know and running on servers you don't control... but don't worry, we don't care to profile you or sell you data... ;)

brk 16y ago
The real violation is that you just drag-dropped your whole iphone backup to a script Yeah, I guess. I just did a quick strings on the backup files, and you could extract a fair bit of info, but it didn't seem like there was much in there that was truly juicy.

schleyfox 16y ago

I'm growing increasingly convinced that the first iphone tracker desktop app reduced precision not so much to "protect privacy", but to cover up exactly how off the data was. I think selling it as though it were your location rather than tower location backs this up. Looking at it in Google Earth it seems to only be accurate when I was on I-95 in New Jersey. I figured the data would be kind of rough, but for a month of my data in NYC it is just off. Apparently I spent most of my time in Bed Stuy and no time in my apartment in Park Slope. It also got Charlottesville, VA confused with Las Vegas. If that actually happened, there must be some great party stories that I haven't been told about yet.

markolson 16y ago
Interestingly, that was my first thought too. I'm the guy who hacked together the javascript which passably reads the database file in-browser for this, and two things stood out to me as I was working: How much data there was, and how not-quite-that-accurate it was. I think the initial app's way of presenting the data has merit because it does hide the swarm of identical points, but it also does make it 'feel' more accurate when you look at it because 'Yes! I live in that circle.' There's sure to be a happy middle ground somewhere.