I have a large dataset (~15,000 household surveys globally) with binary presence/absence data indicating whether farmers practice intercropping (presence) or not (absence), along with GPS coordinates. I want to build a model that best explains the factors associated with the presence of this practice — not necessarily to understand why farmers adopt it, but to identify variables correlated with the practice. Ultimately, my goal is to produce a predictive pixel-level spatial map of intercropping presence.

I’ve seen some studies using spatial interpolation methods (e.g., kriging) but I’m unsure if these are appropriate for a binary response variable. Others mention spatial regression models, but I haven’t found much on their application for binary data like mine.

What would be the best modeling approach for this kind of spatial binary data? Are spatial regression models suitable here, and if so, which ones? Any advice, references, or examples of similar studies would be greatly appreciated!