0
$\begingroup$

I'm very new to this kind of analysis so I hope my question is of good interest. I've GPS locations from three independent GPS loggers, which data consists of two variables: latitude and longitude. Here's a sample of how a dataset coming from GPS logger 1 looks like in R:

> head(dataraw)
   Date & Time [Local]  Latitude Longitude
1:      6/18/2018 3:01 -2.434901  34.85359
2:      6/18/2018 3:06 -2.434598  34.85387
3:      6/18/2018 3:08 -2.434726  34.85382
4:      6/18/2018 3:12 -2.434816  34.85371
5:      6/18/2018 3:16 -2.434613  34.85372
6:      6/18/2018 3:20 -2.434511  34.85376

I've been reading some litterature on how this data can be analyzed since I would like to compare positions accross GPS loggers. It is suggested to test for two things:

1) Normality with Kolmogorov-Smirnov test

2) Homogeneity of variances with Brown-Forsyth test

Now, I have questions on how I should run these tests on R. Should my input data when running the test 1) be the Latitude values from two loggers at the time and test all possible combinations?

ks.test(latitude_logger1,latitude_logger2)
ks.test(latitude_logger2,latitude_logger3)
ks.test(latitude_logger1,latitude_logger3)

And then do the same with Longitude? If there's no normality (i.e. p < 0.05), should I then log transform all Latitude and Longitude values across loggers.

As for running test 2), could anybody put me on the right track?

Any input is appreciated.

$\endgroup$
2
  • $\begingroup$ Reading between the lines, it sounds like you are trying to see if the GPS loggers agree with one another. You've got 3 GPS loggers, and you had each of them record position at the same set of times (hopefully the same). If this is correct, it would probably help to state that as your goal. $\endgroup$
    – Weiwen Ng
    Mar 11, 2019 at 17:37
  • 1
    $\begingroup$ If not, then I'm puzzled why you'd test if the distributions of latitudes and longitudes are normal. Why should they be normal in the first place? Also, per my understanding, the K-S test can be used to test if one sample is normally distributed, but it can also test if two samples have the same distribution. Your question asks about the first case, but your syntax leads us to believe you want to do the second thing. $\endgroup$
    – Weiwen Ng
    Mar 11, 2019 at 17:39

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.