3rd, 4th & 5th clusters and comments about methodology

I am quickly commenting on my final three clusters. The third cluster represents grid squares with fast average response times (under 7 minutes) and low number of immediate incidents in the year (under 150). These grid squares I am guessing are prodeminantely occupied by offices, warehouses and industrial premises - non-residential. The fourth cluster, the grey squares, I think, are grid squares that have mixed premises types in them as they show no extremes. There is obviously a fifth cluster of grid square of no incidents that are shown blank on the maps.

For completeness I should probably mention that the second cluster -blue has a slight negative correlation of -0.30, the third - purple a slight positive correlation of 0.28 and the fourth - grey a very slight negative correlation of -0.13.

There are few more things I want to say. This borough, which I realise is now becoming identifiable, is a borough I never worked in in my 30 year police career so this is an area of London I do not know very well. So my initial analysis is based on the data I have presented and my knowledge of police work and police information systems. The methodology I am using requires that sensible clustering and hypotheses to be made based on logical analysis of the data available. This leads to new lines of research, gathering more data/information to confirm the validity of my decisions or otherwise. This is a revising and refining process which hopefully will end up with a robust clustering and hypothesis.