non-numeric argument to binary operator

[1234 adsf]

Dear Christen,

First of all, thank you for your work to make the ctmm package available.
I am writing to ask your help regarding an error that I ran into while analyzing our GPS tracing data. Normally it works without problems for several individuals of animals, but for one of them I ran into the below error:

ctmm_speed <- speed(telemetry_akde,FIT1_pHREML)
Error in sqrt(diff(data$x)^2 + diff(data$y)^2)/DT/SPD non-numeric argument to binary operator

The data that causes the error was collected with the same method as all the rest that work well with the calculations. For troubleshooting, I have tried two things:

I tried to find the part of code that gives the eror. It might be speed.r line 249 (FRAC <- sqrt(diff(data$x)^2+diff(data$y)^2)/DT / SPD), but it was too complicated for me to track down if it really is, and how exactly the problem occurs.
Another thing I tried was to narrow down the input data to identify the problematic content in it. This attempt was also not successful, as I got a differen error when I eliminated the first half of my data and ran the script. (I saw the progress bar was at 53% when the error happened, so I assumed it might have to do with some data quality issue at about the 53% section of my data.) Whe I did this, I got the "Movement model is fractal" error.

What could be the best troubleshooting steps to solve this?
Thank you!

Máté Pölös

[Christen Fleming]

Hi Máté,

Can you send me a minimal working example (data + script) to reproduce the error?

"Movement model is fractal" isn't an error by itself, but a warning that the data are too coarse to estimate speed.

Best,

Chris

[Matt]

Hey Chris and Máté,

Just wanted to follow up on this and see if either of you had any thoughts. I'm getting the same error while estimating speed.

Thanks,

Matt

[Margaret Mercer]

Hi Chris,

I'm also experiencing the same issue when running the "speed" function. I'll send you a minimal working example.

Thanks,

Margaret

[Christen Fleming]

Thanks Margaret,

I pushed a fix GitHub. Also, with a movement model that tortuous, I would also set robust=TRUE in all of the speed() calculations.

Best,

Chris

[Margaret Mercer]

Hi Chris, 

Thank you! It's taken a while but I finally got the speed() function to work, although I'm getting Inf for many of my individuals. I'll troubleshoot it myself and reach out again if the problem persists.