I am looking to assess the correlation of three specific variables on a global scale. The data is currently a gridded shapefile, each grid cell with a value for all three variables. Obviously, something like a regular Spearman Rank correlation isn't helpful here. Is Moran's I appropriate here if I want to be able to say "these three variables are correlated / not correlated globally and/or variable 1 is correlated with variable 2 but not variable 3." If so, what is the best way to calculate Moran's I when you have a gridded shapefile. If not, what is a better approach and how would you implement it? Thanks for your consideration!
Hello,
I think the best R package for univariate and correlogram plots is spdep with Moran's method (https://www.rdocumentation.org/packages/spdep/versions/0.8-1/topics/sp.correlogram). If you want to explore your data through a multivariate analysis, I think you can use mpmcorrelogram function (Multivariate Partial Mantel Correlogram, https://rdrr.io/cran/mpmcorrelogram/) or mantel.correlog (vegan package, https://www.rdocumentation.org/packages/vegan/versions/2.4-2/topics/mantel.correlog). I used this last option once. You can adjust your data to the provided example.
I hope this helps!
Cheers,
IM
- Badges
- Science topic
- Similar topics
- Political Science
- International History and Politics
- Globalization