U.S. flag

An official website of the United States government, Department of Justice.

Forecasting Municipality Crime Counts within a Metropolitan Area

Award Information

Award #
2009-IJ-CX-0026
Funding Category
Competitive
Location
Awardee County
Philadelphia
Congressional District
Status
Closed
Funding First Awarded
2009
Total funding (to date)
$497,981

Description of original award (Fiscal Year 2009, $497,981)

This award will provide funding to Temple University to build crime models with temporally lagged predictors to forecast later out-of-sample, one- and two-year-ahead crime counts in 353 municipalities in the country's fifth largest metropolitan area, Philadelphia-Camden. The work will examine several threads of theoretical relevance in the communities and crime literature. These will include but not be limited to understanding why crime counts in some municipalities might go up or down faster or slower than the area-wide average, and whether positionally and
historically determined municipality type is more important than recent demographic indicators in shaping municipality-level crime trends. Two indices of crime counts, for property and for violent crimes, will be the outcomes of interest in raw form, differenced form, and population weighted percentile form. The proposed project is economical because it builds on an already-existing infrastructure for gathering, tabling, analyzing, and making available a wide set of metropolitan indicators. The Temple University Metropolitan Philadelphia Indicators Project (MPIP) has assembled yearly Municipality-level crime municipal indicators, made them available through a web front end permitting mapping and statistical table access, and has built strong relationships with local, regional, and state leaders and policy makers. The proposed project updates all indicators, fills in and updates crime data, adds yearly demographic estimates for 2001-2008, adds municipalities within three jurisdictions of the metropolitan area boundary in order to accomplish spatial smoothing, builds models of crime counts and crime changes, and uses the best models to forecast one-year-ahead municipality-level crime counts and crime changes, and two-year-ahead crime counts. Law enforcement census indicators are used as potential explanations of forecast errors.

ca/ncf

Date Created: September 22, 2009