Stack Exchange network consists of 183 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.
@wingnut I researched that library earlier and it continuous. I was able to generate Delaunay triangulation for my polygon data after converting them to centroids. But i don't know how to finish it in the form of spatial weight matrix.
@Spacedman Islands and disconnected regions are two different things in PySAL. I computed Global Moran's I with queen contiguity. But in case of Local Moran's I, PySAL is not allowing weights with islands but disconnected components are fine here. Anyways I calculated weights for my dataset by using Kernel weights with adaptive distance like weights.distance.Kernel.from_dataframe(gdf, fixed=False, k=15) and that worked fine. But for weights with islands like Queen, PySAL's Local Moran I is not working and In documentation, no instructed for not using queen or rook weights.
@Spacedman If we use KNN weights then Moran's I can be calculated for this same dataset. I'm thinking that maybe it is not possible to calculate Moran's I for the datasets with islands using Queen or Rook` weights.
@Llaves This answer worked well but I don't know why this geopanda's developers keep things complicated. The process I shown should have worked without any hassle.
@Jake Shermeyer I agree that data management for VIIRS is worse but the grid is correct. They designed a new grid for VNP46 product for some good reason. You can find it in VNP46 user's guide.
