I'd like to begin by introducing a computer game uniquely designed to illustrate the concepts and techniques used in more serious real-world crime-related applications.
SimCity, for those who haven't had the pleasure, is a popular computer game that lets players build and manage a city. Several "layers" (or themes, as they're sometimes called) of spatial information are available to help the player determine what course of action to take, depending on the player's analysis of the data. Available game themes include traffic flow rates, pollution levels, and criminal activity, to name a few. The player acts as the mayor to make decisions on how to deal with the city's problems, using these layers and the recommendations of advisors to determine and implement correct policies.
Like the game, real-world businesses or government departments rely on the spatial location of several entities to perform their functions. Transportation managers analyze traffic patterns to determine the performance of the system, environmental or pollution experts make recommendations to help improve air and water quality, and law enforcement officials develop strategies to fight crime.
Everyone is familiar with scenes in TV detective shows where detectives use a map and colored tacks to show the location of crimes with similar characteristics, or modus operandi. The detectives analyze the map and the points representing crimes, trying to find out where the next crime might occur, which shop owners police officers should interview, or where the suspect might be hiding. This type of point data is an example of a vector dataset. Other information can be generated in vector format, with representations in line or polygon format. For example, transportation engineers can represent traffic flow as a line with direction and magnitudethe direction of traffic and the volume of vehicles over time. (And you thought you'd never use high school physics again!)
In SimCity, one of the available data layers is a thematic map of criminal activities. The density of criminal activity is mapped, somewhat like a weather map on the Weather Channel, with areas of high crime rates drawn or colored in dark shades, and areas of lower crime rates in progressively lighter shades of the same color. The data used in the game is an example of a raster dataset, which is a grid or cell-based dataset in which each cell contains a value. Thematic maps can also be created as vector datasets. Figure 1 shows two thematic maps, the first a vector population-density map of Colorado using polygon data, and the second a fictional raster thematic map of the crime rate for an area. In the top map, population densities for each county in the state range from high (dark orange) to low (light orange). In the bottom map, high-crime cells are represented with dark red, with decreasing rates represented by progressively lighter shades. These are among the simplest kinds of analyses and displays that can be doneresearchers continually come up with more interesting and complex methods to help solve more sophisticated problems.
Sample thematic maps.
Real-world law enforcement officials also have a wealth of spatial and associated attribute data available to help fight crime. These data layers would generally be created to help solve problems on one or several different scales, or, as I like to think of it, relative "areal" extents. A high-level official, such as a mayor, would generally be interested in the crime rate for the entire city, which is a large-scale issue, while a district police chief would be interested in the criminal activity within his or her district boundaries, a smaller-scale issue. Those who deal with water pollution are often interested in scales ranging from small scale, such as the Red River watershed, to large scale, such as the Great Lakes basin.
After determining that the city can afford it, questions such as "Where should we build a new police station to fight the crime problem?" or "What neighborhoods would benefit the most from increased funding and support?" can be answered, in part, by analyzing available GIS data. Subsequent mappings and analyses can help determine whether the optimal location was chosen, if further funding to the station is required, and so on.