Statistical tests for hot spots
A primary step that identifies whether hotspots exist in geocoded crime and disorder
data are global statistical tests for clustering and dispersion. This first step
can provide insight into what types of patterns will be expected when the crime or
disorder data are mapped. For example, it is often assumed that crime and disorder
data is concentrated at certain locations. Preliminary global tests can confirm whether
there is statistical evidence of clusters (i.e. hotspots) in crime and disorder data
which can then go on to be mapped, rather than mapping data as a first step and struggling
to accurately identify hotspots when none actually exist.
Two groups of tests can be applied depending on the precision of the partnership’s
data. Click below for further information on the two groups of tests.
CrimeStat.
A suitable and freely available application to use to run these tests is a product
called CrimeStat. CrimeStat is a United States Department of Justice product, supported
by the Crime Mapping Research Centre. The program is Windows-based and interfaces with most desktop GIS programs. Its
purpose is to provide supplemental statistical tools to aid geographical crime analysis.
CrimeStat offers an easy way to test for evidence of hotspots and relative levels
of dispersion. CrimeStat also offers many other analytical routines, many of which
are probably too technical for typical partnership requirements, tend to apply better
with American data, or can be performed using standard GIS software tools.
|
