.
(2) Standard Deviation Ellipses: Regions based on standard deviation are often used in crime analysis in predictive models for serial crimes. If an analyst has determined that a crime was likely committed by a single person or group, known crime locations can be used to help predict where and for what purpose law enforcement personnel can be deployed (Nulph, Burka and Mudd, 1997). .
(3) Change Maps: Maps can be generated which show where crime is staying the same/increasing/decreasing between any two independently selected periods of time. Histograms can also be produced displaying the time of day and day of week when crimes occurred. Crime data is, however, rarely definable as one particular moment in time. Victims of a burglary might go away on holiday and return a week later to find they have been the victims of a break-in, and unable to tie the time of the incident down to a specific moment. Recent work in Aoristic Crime Analysis hopes to solve this problem (see Ratcliffe and McCullagh, 1998). .
(4) Linkages: In many cases, the data examined in crime analysis is better considered as dyads, rather than as single point features. The criminals home may have a bearing on where a crime is committed. A car may be stolen at one site and recovered at another. If these connections are not considered some patterns of behaviour may become obscured (Nulph, Burka and Mudd, 1997). .
The second level of crime analysis focuses on individual crimes. The GIS program allows the user to search for crimes of a particular nature, perhaps because a particular suspect is being targeted or because they have a distinctive profile. Attributes can be assigned to each pin', which can be bought up on-screen simply by clicking on them. Whereas physical pins rarely have more than one attribute (identified by their colour), computerised pins' can have unlimited attributes, and act as a window onto an underlying database.