Efficient Variography with Partition Variograms

I am happy to share a paper accepted in C&G in which we generalize the concept of directional variograms to general partitions of spatial data. In the paper, we illustrate this new concept of partition variograms, and provide a parallel algorithm to efficiently estimate spatial correlation in large data sets of arbitrary configuration. The manuscript is available online, and the implementation is part of the GeoStats.jl project.