0
$\begingroup$

I have a dependent variable which is tree cover loss and independent variable which is temperature. Both variables have 20 groups/regions in a country in them and I want to do some kind of stats test/correlation test but am not sure which one to do?

I have each region with the years from 2001-2021 but I also have just the mean of each region if I just need the mean.

Improve this question
$\endgroup$
4
  • $\begingroup$ A linear model! (maybe) $\endgroup$
    – user2974951
    Commented Feb 17, 2023 at 11:57
  • 1
    $\begingroup$ 0. Welcome to CV.SE. 1. What is the distribution of the "tree-cover loss" variable like? In general, the standard "Data Science Process" applies. 1. "Business understanding" (identify our objectives, metrics, etc)." 2. "Data acquisition and understanding" (get the necessary data and do some EDA on them), 3. "Model" (what is being asked), 4. "Deploy" (not very relevant here), 5. "Customer acceptance" (e.g. publication? good grade? etc.) We are still in points 1&2 here (i.e. what are the data at hand). :) $\endgroup$
    – usεr11852
    Commented Feb 17, 2023 at 13:09
  • $\begingroup$ Show us some plots? A link to (excerpts of) the data? $\endgroup$
    – kjetil b halvorsen ♦
    Commented Feb 19, 2023 at 14:29
  • $\begingroup$ Because you haven't specified a test before collecting the data, it sounds like you are trying to discover something about the relationship. Testing is not the way to do that. Instead, view graphical representations of the data. A scatterplot would be a good start. $\endgroup$
    – whuber ♦
    Commented Feb 19, 2023 at 14:38

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.