Hi I have read many of the posts regardng Morans I and Hot Spot and after reading a couple of other sources have a question which would like to discuss: The resource centre video regarding the Hot Spot Analysis (one with the obesity rates) stated that using Global Morans I could help us determine an appropriate distance band by finding a local maximum of Z score (meaning 'a tip' if you use a graphic). Then I found a paper by Pei Fen Kuo, Xiaosi Zeng and Dominique Lord of the Texas A&M University named Hot Spot Guidelines (https://ceprofs.civil.tamu.edu/dlord/Papers/Kuo_et_al._GIS_Guidelines.pdf). In this paper using a small scale example of data in a 7/5 matrix it was shown that Hot Spot analysis tends to miss single 'hot values'. This means that Hot Spot is more reliable with clustered than with dispersed data. Now, the Z score in Moran's I is a probability score, while the I score is a measure of clustering/dispersion the question is: Isn't finding the I score's 'tip' more appropriate for a better distance band than the Z score in Morans. Or at least a combination of the two. But only looking at the Z score, given the assumption that clustering/dispersion (I score) has a great effect upon hot spot, is a bit illogical to me. I apologize if an answer was given somewhere or an answer could be deduced from an other post. I tried to look thoroughly Ivan
