A set of sample maps have been assembled to demonstrate how crime maps can be used. For more information on the terms used in these examples, please read Crime Mapping Using Incident-Based Data
The first three maps illustrate the use of UCR summary data to produce choropleth maps. As the level of focus or scale is changed, new perspectives on crime appear; this shift illustrates what is called the "cone of resoultion".
This map uses national 1999 crime data to show violent crime rates.
The next map also uses 1999 crime data to show crime at the county level in Massachusetts.
The third map uses 2000 crime data collected by the Massachusetts Crime Reporting Unit to show violent crime rates at the city level.
The next map is based on address data collected as part of the Massachusetts enhanced NIBRS system, which allows for the collection of street information. This permits the aggregation of crime data into real units that are more specific than would be available using standard UCR or NIBRS data.
This map shows violent crimes near Worcester by Census block.
The next set of maps shows the rate of victimization for assault offenses (aggravated assault, simple assault, and intimidation) in residences in Worcester by Census block groups in the first quarter of 1996.
The first map shows the location of all assault offenses occurring in residences.
The next shows only those assaults with male victims.
The final map in this set shows only those assaults with female victims.
The next set of maps demonstrates the importance of checking the "hit rate" of your data before too much additional analysis is done. For most purposes, a hit rate of 90% or greater is needed in order to get a good picture of the data in question. This point becomes especially important when attempting to develop an interagency analysis capability. If one agency has a significantly different success rate than other agencies, it would be misleading and potentially dangerous to compare crime patterns across the two jurisdictions. It would be important as well to analyze the types of incidents that are not geocoded to determine whether these addresses are clustered in one part of the jurisdiction and to determine whether the frequency distribution of these offenses are significantly different than the geocoded incidents.
The first map shows the success rate for geocoding crime data across the state of Massachusetts.
The second map shows the success rate in geocoding crime data for the area around Worcester.
The final map shows an analysis of marijuana and heroin arrests in one city. The map illustrates two key points. First, arrest locations are marked by variably sized points or dots; the size of these points is proportional to the amount of drugs seized. Second, using CrimeStat II, a program written by Ned Levine and sponsored by the National Institute of Justice, it is possible to do a "nearest neighbor" analysis, which allows the analysis to determine standard deviation ellipses around clusters of points. As seen in the map, this analysis indicates that the pattern of arrests for the two drugs, although overlapping in part, are different. This might suggest that different enforcement practices are needed to address the drug problem in this community. It might also suggest that there is not one drug problem, but rather a series of drug problems.