Fig. 3

The first 2018 rows of the Seattle Police Department data are used (taken from data.gov), and the results can be compared with those of Fig. 2 where the full set of events are used. Subfigure a shows the cluster entropy trace for the window size set. There is greater variability in the trace of the cluster entropy variance estimation for the larger window size of 100 which shows that more data is required to ensure similarities in the subsetting positions and that smaller window sizes have more cross comparisons to reduce the variability. In subfigures b and c it can be seen that crimes ’ARSON’, ’THEFT-BICYCLE’ and ’RAPE’ are not present in the subset data of these events. It can also be seen how the crime ’NARCOTIC’ is recorded in ’Precinct WEST’ where in the previous examination it was not