Nearest neighbor ordering algorithms (hierarchical clustering, dendrograms) are well suited for sorting the items in either or both axes. If either axis relates

Nearest neighbor ordering algorithms (hierarchical clustering, dendrograms) are well suited for sorting the items in either or both axes. If either axis relates directly to a physical dimension then it might be best to maintain that order in the matrix. A lot of the fun is in working out how to measure distance between items based on their values as that drives the results of the nearest neighbor order.