1

I have spatial temperature data on a relative scale from 1 to 0 on several days. The data sets have the same resolution, origin, and extent.

How can I statistically compare them? How to identify patterns? What measures could be used to give a value for overall similarity?

The mapped data looks like this:

enter image description here enter image description here

mkt
  • 18,245
  • 11
  • 73
  • 172
Lipsan
  • 11
  • 1
    If oyu flatten the values for each map into a single vector. So in your example you would have two vectors. You could compute the cosine similarity between the two vectors. – Janosch May 18 '22 at 13:06
  • There are lots of possible metrics for comparison. To narrow them down to ones that are good for your purposes, consider a pattern shifted north by one pixel, or shifted colder one degree, or rotated 90 degrees. Of those, which should be considered most similar and which most different from the original pattern? – Matt F. May 19 '22 at 06:41
  • 1
    These maps look like the results of interpolations. Thus, any comparison is likely to tell you as much or more about the interpolation method than about the data. Could you therefore describe the original data from which these maps are derived? When you make those edits, please narrow your questions. E.g., how to "identify patterns" is too vague and broad to be answerable here on CV. If that's not evident, take a look at https://gis.stackexchange.com/a/56573/664 for a tiny glimpse of just some of the considerations involved in a similar situation. – whuber May 19 '22 at 13:36

0 Answers0