U.S. flag

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

NCJRS Virtual Library

The Virtual Library houses over 235,000 criminal justice resources, including all known OJP works.
Click here to search the NCJRS Virtual Library

Improving Geocoding Rates in Preparation for Crime Data Analysis

NCJ Number
218549
Journal
International Journal of Police Science & Management Volume: 9 Issue: 1 Dated: Spring 2007 Pages: 80-92
Author(s)
Allan J. Brimicombe; Lily C. Brimicombe; Yang Li
Date Published
2007
Length
13 pages
Annotation
This paper presents an improved automated approach to increasing the mapping rate for crime data, so as to make crime analyses more useful.
Abstract
The new geocoding toolkit (matching a crime to the geographic location where it occurred) has been developed in order to improve the "hit" rate (rate at which a batch of crimes can be accurately located on a map). The purpose of the toolkit is not to replace commercial address-matching software such as Matchcode or QAS, but to enhance the outcome of the geocoding process by building additional steps and tools around these existing software products. It is a five-stage process. The first stage cleans common errors that arise in the address fields of crime data. In the second stage, the crime data are passed through commercial address-matching software, which attaches geographic coordinates to the crime location based on a street address. All addresses successfully geocoded at this stage are given the validation code "L1," indicating that the crime has been linked to an individual property address at the highest level of accuracy. The third stage focuses on crimes with nonaddress locations. The majority of these are street junctions and can be found in the free-text data field that describes the venue of the crime incident. Other nonaddress locations would include railway stations, bus stations, and prominent landmarks. The junctions are text-mined by searching for key words. In the fourth stage, all remaining records with a valid unit postcode (mail delivery point) are geocoded at the postcode level. The final stage of the toolkit geocodes all remaining records according to street name. A test of this system in a British police force raised the "hit" rate for accurate crime location an additional 65 percent to a rate of 91 percent.