|| Spatial statistics program for the analysis of crime incident locations
|| The program inputs point (incident) locations and calculates various spatial statistics. Input can be in spherical or projected coordinates and the routine will calculate a reference grid. Spatial interaction can be measured by either distance (direct or Manhattan) or through a travel network which allows the interaction to be measured by time, speed, or travel cost. The spatial statistics in CrimeStat are subdivided into seven major categories:
1. Statistics describing the spatial distribution of incidents (such as the mean center, center of minimum distance, standard deviational ellipse, Moran's I spatial autocorrelation index, Moran correlogram, convex hull, or directional mean).
2. Statistics for describing properties of distances between incidents - nearest neighbor analysis, linear nearest neighbor analysis, Ripley's K, the assignment of primary points to secondary zones based on either nearest neighbor or point-in-polygon, and distance matrices calculations.
3. 'Hot spot' analysis - mode, fuzzy mode, hierarchical nearest neighbor clustering, risk-adjusted hierarchical nearest neighbor, STAC, K-means clustering, and local Moran statistics.
4. Single-variable kernel density estimation routine for producing a surface or contour estimate of the density of incidents and a dual-variable kernel density estimation routine for comparing the density of incidents to the density of an underlying baseline.
5. Space-time analysis - Knox index, Mantel index, spatial-temporal moving average, and correlated walk analysis.
6. Journey-to-crime analysis including calibration, estimation, and the drawing of crime trips. The journey-to-crime function can be modeled with five different mathematical functions or an empirically-derived function.
7. Crime travel demand, an application of travel demand theory to crime analysis. The model is conducted at the zonal level and includes modules for:
- Trip generation - separate models for trip productions and trip attractions using a stepwise, multivariate Poisson regression model with an overdispersion correction and a balancing procedure for ensuring that productions equals attractions.
- Trip distribution - calculating the observed trip distribution, modeling the trip distribution, and comparing the observed with the predicted trip length distribution. Travel impedance can be modeled with five different mathematical functions or an empirically-derived function.
- Mode split - estimating the likely travel model split for each origin-destination pair for up to five different travel modes.
- Network assignment - estimating the likely travel routes taken on a travel network including the total volume on each network segment. The network can be modeled using travel time, travel speed, or travel cost in addition to distance. Several network utilities are provided.
|| Not Available
|| The program is accompanied by sample data sets and a well written manual which gives the background behind the statistics and worked examples. Examples from researchers in different fields are presented throughout the manual
CrimeStat writes graphical objects to ArcView, MapInfo, Atlas*GISTM, Surfer for Windows, and ArcView Spatial Analyst.
Version 3 (November 2004) with updates to the manual through May 2005
Information about crime mapping can be found at www.ojp.usdoj.gov/nij/maps/
|| Home: www.icpsr.umich.edu/CRIMESTAT/
Ned Levine, CrimeStat: A Spatial Statistics Program for the Analysis of Crime Incident Locations. Ned Levine & Associates, Houston, TX and the National Institute of Justice, Washington, DC. November 2004.