Rather than pick up no data, which would have been preferable because the system interpolates from the surrounding data if it goes through a patch of bad reception, it actually picked up wrong data, which is way harder to correct for, in this instance anyway.
So, here (actual route) and here (GPS lies) we have two maps, one with the trace from the GPS, one with the actual route I ran. They are the same pixel size and more or less the same scale.
And here we have a (quite large) Photoshop file with one overlayed on the other, the top one at 50% opacity, which might help get a handle on it a bit better if you switch the layers on and off a bit.
And here we have the GPS data in the form of an XML file.
I have struggled with this for a few hours now and I'm not making much in the way of progress.
If it was just a training run, I'd not be bothered, but I want to keep the marathon data for posterity, and I want it to be right.
Can anyone help me munge the XML so that the trace resembles the actual route run?
It involves working out which points are responsible for the erroneous plots, and correcting them with sensibly-interpolated values from the good data either side.
Sounds easy, doesn't it.
A bottle of fine wine, or a toy, or something -- an as yet undecided present -- awaits the first person who can do it. Not to mention my undying gratitude.
If you need to test the new XML once changed, either I will need to do it, or else you'll need to steal the code from my run tracker page (http://ure.org.uk/diary/runmaps/workoutTracker.php, also there as workoutTracker.php.txt, and workoutTracker.js is there as workoutTracker.js.txt) and get yourself a Google Maps API key (free, and easy, from google). Unless there's some other way of testing it.