Hi all! I’m analyzing high resolution satellite telemetry data (GPS positions and accelerometry data) recorded by means of data loggers attached to animals (mainly birds). I’m testing several r-packages to estimate home range, space use and movement parameters (i.e., speed, turn, ODBA, etc.) including the “adehabitat” family (adehabitatHR, adehabitatLT, adehabitatMA), “move”, “BBMM” and “rhr”. The problem is that most of these packages have problems (r stops computing, crash, etc.) when large datasets (> 0.5 million locations per animal) are analysed.
Do you know any r-package appropriate for these kinds of datasets in which data are extremely autocorrelated (locations recorded at 1 sec interval)?
Thanks
Pascual López, Ph.D.
University of Alicante, Spain
https://paslopez.wordpress.com/
https://paslopez.wordpress.com/
Hi Pascual,
I'm surprised there have not been any answers yet so I thought I'd try to get the ball rolling by sharing my experience. I just had a paper published where we had 1hz GPS data on 25 individual baboons at once and we were interested in their relative movement and how they make collective decisions. First thing is that R < 3.0 could not cope with this data. In 3.0, they changed some of the architecture which made a huge difference. We can now easily run vectorised functions on all animals at once (15 million data points per animal). However to do this you will need a lot of RAM (I have 16gb). I'd say that this is most likely to be the stumbling block for you as if there isn't enough RAM available it will appear as though R has crashed.
My second point is that, for much of what we wanted to do, it was easier to make our own code. For example ODBA is pretty easy to calculate (and I recommend considering using VDBA instead). Also depending on your questions you might find it helpful to subsample your data. For example home-range calculations won't change much if you only take 1 point per minute.
Good luck with this. Its not easy but it sounds like you have some really interesting data that will be very rich.
Damien
Hi Damien,
Thank you very much for your kind response. It's really useful. I have downloaded your paper and will contact you privately if I have any particular question about methods. Congratulations for your work and for your publication! It's amazing!
All the best,
Pascual
Hi...Similar problem I faced. Upgrade your RAM in PC. Hope it will work...good luck..
- Badges
- Science topic
- Similar topics
- Filing