- Converting .mov files to .avi in Microsoft Windows
- Anne Doyle
- Fields of Athenry 10k - 2011
- Michael D and The Mountain
- Short walks in Northumberland
- Testing optimisation algorithms: Part IV - Objective function and optimisation in Python
- Testing optimisation algorithms: Part III - Global optimisation in R
- Testing optimisation algorithms: Part II - Implementing objective function in R
- Testing optimisation algorithms: Part I - A circle packing objective function
- Oscillations
Parsing and analysis of GPX files in R
Sat, 05/14/2011 - 22:38 — Conor
As I mentioned in a previous post I've been working on ways to parse the .gpx XML files which are generated by my Garmin Forerunner so that I can analyse the captured data more carefully. The online data visualisation tool provided by Garmin is good, but there are a few flaws. It's not always possible to scale data nicely, it's not possible to compare between datasets, and a few other things. Reading the raw data into R would allow all kinds of analysis.
R is a very powerful, flexible, open-source statistical programming language which is really good for parsing and analysing datasets.
I've written some R code to parse .gpx files, analyze the data and generate .pdf reports. You just put the script into a directory, together with the .gpx file to be analyzed (which you can access by clicking the export button on the page for any of these activities for example), add the filename to the top of the script, and run it. A .pdf report like this one is generated.
The reports contain several types of plot, including the following four types:
Elevation varying along route followed. Colour indicates elevation, x and y coordinates represent distance from the starting point in a northerly and easterly direction respectively.
Heart rate varying along route followed. Colour indicates heart rate, x and y coordinates represent distance from the starting point in a northerly and easterly direction respectively.
Pace varying along route followed. Colour indicates pace, x and y coordinates represent distance from the starting point in a northerly and easterly direction respectively.
Elevation, heart rate, speed and pace varying with distance covered along route.
The R code is effective but untidy. I had intended to write it nicely into S3 classes or even an R package, as I think that others might be interested in using it, but I have run out of steam a bit. If anyone is interested in it, let me know as this would give me some motivation!
Here are some other reports I've generated. The first four runs are along essentially the same route and are comparable.
- Conor's blog
- Login or register to post comments
Comments
Update for normal .GPX files
Peter Smith pointed out that this script does not parse non-Garmin .gpx files correctly. This was due to a missing heart-rate element in the .xml for normal .gpx files and the near certainty that I am parsing the .xml in a horribly clunky fashion.
I've created a second version of this script which should work for non-Garmin .gpx files (at least it works for the example file that Peter kindly sent to me). This is just as well as, to my dismay, my Garmin device packed it in a couple of weeks ago. I hope they'll replace it for me, otherwise I will be looking out for alternative devices and data storage methods.
The R script for pure .GPX files can be downloaded here.
As with the other script, place the script in a directory with your .gpx file, edit the filename at the head of the file to match your filename.gpx, start an R session, set your R working directory to the one with the script and the .gpx file in it and run the script. It will generate a .pdf report.