Compare field characteristics (e.g. area, shape metrics) between countries or between years
Sampling problems:
Come up hypothetical crop type distributions and sample scheme to collect from road network - evaluate sample-based estimates relative to true distribution
Modelling demo
Random Forests
# Train modellibrary(caTools)library(randomForest)library(geospaar)# read in train settrain_ref <-system.file("extdata/train_reference.csv", package ="geospaar") %>%read_csv() %>%mutate(crop =factor(case_when( class =="Maize"~1, class =="Rice"~2, class =="Other"~3, class =="noncrop"~4))) %>% dplyr::select(id, class, crop, !!names(.))