I am working on a landscape-scale spatial analysis of fire behavior using mlr3 and as a follow-up to the formal analysis where I plan to report global variable importance, I would like to examine the local variable importance to qualitatively examine how the importance of certain landscape variables changes spatially. This was possible in mlr by setting the argument local=TRUE when using the [generateFeatureImportanceData] function, but I cannot seem to find an equivalent in mlr3. There appears to be a variety of options for examining local importance for single observations, but nothing for examining an entire datasets (i.e., landscapes). Is there an mlr3 equivalent to ‘local = TRUE’, or is reverting to the original mlr package for this functionality my best option? Many thanks in advance.