MAPS Projects: Software and Technology
- Development of a Web-Based Geographic Information System for the National Missing and Unidentified Persons System (NamUs)
- The Expansion of the CrimeStat Spatial Statistics Program to Incorporate Spatial Regression, Space-Time Clustering, Bayesian Journey-to-Crime Modeling Incident Detection Methodology, Sparse Matrix Tools and the Implementation of Recommended Changes
- The Mathematics of Geographic Profiling
- The Restructuring of the CrimeStat III Spatial Statistics Software Program
- Three-Dimensional Extensions to Geocoding Technology for Emergency Response Applications
- Urban Growth Simulation and the Changes in Crime Patterns
- Using Open-Source Technology to Enhance Post-Release Supervision Systems
Development of a Web-Based Geographic Information System for the National Missing and Unidentified Persons System (NamUs)
| Grantee: | University of Tennessee |
| Amount: | $165,509 |
| Award Number: | 2008-91574-TN-IJ |
Description: This proposal requests funding to support the development of a Web-based geographic information systems (GIS) for NIJ's National Missing and Unidentified Persons System (NamUs). The proposed system will allow law enforcement, medical examiners, forensic anthropologists and other stakeholders to spatially query both the missing persons and unidentified decedent databases. The graphical user interface (GUI) will allow the user to identify a point location, linear feature or area to base the query.
The Web application will tie directly into the NamUs system, and the application will enhance the search capabilities of NamUs. These improvements should provide stakeholders with a more focused list of potential missing persons or unidentified decedent cases. Dr. Ralston, with support from the Department of Geography at the University of Tennessee, will coordinate the development of the Web-based GIS application. A researcher at the Forensic Anthropology Center will coordinate data checking and geocoding of unidentified decedent reporting system records and missing persons cases in NamUs. These efforts will require communication and travel by a representative of the University of Tennessee's Forensic Anthropology Center to regional forensic centers, coroners' offices and morgues to update NamUs records. The funding will also support a joint meeting of the GIS applications development working group with select potential stakeholders from previous NamUs workshops to test beta versions of the application and to modify the GUI.
The Expansion of the CrimeStat Spatial Statistics Program to Incorporate Spatial Regression, Space-Time Clustering, Bayesian Journey-to-Crime Modeling Incident Detection Methodology, Sparse Matrix Tools and the Implementation of Recommended Changes
| Grantee: | Ned Levine & Associates |
| Amount: | $537,790 |
| Award Number: | 2005-IJ-CX-K037 |
Description: This grant expanded the CrimeStat spatial statistics software program to meet the growing need for technology that can detect emerging crime hot spots in real time. This research developed a new version of CrimeStat that incorporates incident detection routines that can be analyzed at different levels of spatial aggregation (e.g., the block level, the neighborhood level and the district level). It incorporated improvements in spatial regression modeling to handle very sparse data sets. In addition, the research will create CrimeStat "building blocks" that eventually can be strung together to allow wide-area spatial querying and analysis. In the future, officers will be able to remotely query large spatial databases using Web servers, and CrimeStat will be an important tool in the analysis of spatial trends.
The research will improve CrimeStat by (1) incorporating incident detection routines into CrimeStat to provide monitoring of events in both space and time (these routines can now be analyzed using different levels of spatial aggregation); (2) expanding spatial regression modeling to include very sparse matrices, again using spatial aggregation techniques; (3) developing a version for network and Web-server environments using the Microsoft Windows Server 2003 operating system to allow Web-based applications using CrimeStat, and a version for 64-bit computers using Microsoft Windows XP Professional x64 Edition to allow the analysis of very small zones (e.g., blocks); (4) developing a test for the tools with data from Chicago and Baltimore; and (5) creating a prototype training course on the routines at the Baltimore County and Chicago police departments.
The research team included some of the leading world authorities in crime mapping and spatial statistics. Ned Levine, an expert in safety and security modeling, developed the CrimeStat program for the National Institute of Justice (NIJ). Shaw-Pin Miaou, a statistician, has developed statistical tests and spatial regression techniques using the Bayesian approach and is very knowledgeable about incident detection technology. Richard Block is a criminologist who was one of the original developers of the Storage Allocation and Coding Program. Phil Canter, a crime analyst at the Baltimore County Police Department, has been instrumental in developing crime mapping applications and in making those applications widely available to the field of law enforcement.
The Mathematics of Geographic Profiling
| Grantee: | Towson University |
| Amount: | $165,173 |
| Award Number: | 2007-DE-BX-K005 |
Description: Geographic profiling uses information about the locations of crimes committed by a serial offender to estimate the location of that offender's anchor point. Currently, three software systems — CrimeStat, Dragnet and Rigel — are used for geographic profiling. Although the details of the approaches taken by these software packages differ, they share a common mathematical heritage.
The purpose of this project is to develop a fundamentally new mathematical framework for geographic profiling. This new framework should (1) be mathematically rigorous; (2) have explicit connections between assumptions on offender behavior and the components of the mathematical model; (3) take into account local geographic features, in particular, geographic features that influence the selection of a crime site and the potential anchor points of offenders; (4) be based on data that are local to the jurisdiction(s) where the offenses occur; and (5) return a prioritized search area for law enforcement officers.
Researchers propose to develop an algorithm based on Bayesian statistical techniques that fulfills these goals. The method begins with the explicit assumption that the offender's choice of targets depends only on (1) the distance between the target and the offender's anchor point and (2) the local geographic features of the target location.
The algorithm then requires a representative list of historical crimes of the same type as the series; this is used to estimate the local target density. The algorithm also requires an estimate of the prior distribution of offender anchor points. This can simply be constant (a noninformative prior); another option would be to use local population density as a proxy for the prior distribution of anchor points. The precise choices of both the historical data set and the prior are left in the hands of the analyst. Finally, the algorithm requires an estimate of the distance decay function of a typical offender (this information is already commonly used by the existing geographic profiling techniques).
Researchers will develop a mathematical algorithm based on Bayesian methods to estimate the probability density surface for the offender's anchor point. Once the mathematical modeling is complete, researchers will produce prototype software that implements the algorithm. The software will be available at no cost to researchers and law enforcement agencies.
The Restructuring of the CrimeStat III Spatial Statistics Software Program
| Grantee: | Ned Levine & Associates |
| Award Number: | 2005-IJ-CX-K037 |
| Intramural Researcher: |
Ronald E. Wilson |
Description: During the past decade, practitioners and researchers have relied increasingly on the CrimeStat spatial statistics software program to analyze crime data. The original program, developed in 1997, had only a limited capability to interact with other software or systems. Today, versatility is instrumental to the success of any software program. NIJ will work to revise CrimeStat so that it can be directly integrated into law enforcement and criminal justice systems. MAPS program researchers are developing a prototype with a more accessible interface, a complete statistical analysis package and a database of information analysts can use to create software programs specific to their jurisdiction's needs.
Three-Dimensional Extensions to Geocoding Technology for Emergency Response Applications
| Grantee: | North Carolina State University |
| Amount: | $564,304 |
| Award Number: | 2007-DE-BX-K010 |
Description: This research study is developing a three-dimensional (3D) address geocoding technology for use in large buildings. Existing two-dimensional (2D) systems have shortcomings in emergency responses involving large buildings, where more detailed information regarding point-of-entry (PoE) and indoor routes is required. Responders need route maps and directions not only to the building PoE, but also from the PoE to the site of the incident using the building's hallways, elevators and stairways.
This project proposes using a 3D geometric network model for structures that is a direct extension of the 2D model used for street address matching and routing. The 3D model simplifies indoor details by representing corridors using a straight medial axis transformation and a 3D geometric duality to represent connectivity between horizontal and vertical paths. A 3D extension to the 2D geocoding process will be developed to provide indoor-route and map-generation capabilities.
Urban Growth Simulation and the Changes in Crime Patterns
| Intramural Researchers: |
Ronald E. Wilson, Derek J. Paulsen, Jay Lee |
| Award Number: | 2007-IJ-CX-K014 |
Description: Many urban planners anticipate and address problems associated with urban growth by using computer programs to predict growth in and around cities or towns. These programs are called urban growth simulators (UGSs). This research will develop a new UGS to analyze which neighborhoods in an urban area are most likely to experience an increase in crime as a result of changes in urban land use (growth, redevelopment and sprawl).
Using Open-Source Technology to Enhance Post-Release Supervision Systems
| Grantee: | The Providence Plan and The Urban Institute |
| Amount: | $273,000 |
| Award Number: | 2007-IJ-CX-K021 |
Description:The Providence Plan and The Urban Institute requested funding from NIJ to develop a geospatial application designed to help corrections, public safety and social service agencies better supervise and assist returning prisoners. In the proposed project, The Providence Plan will design a Web-based tool to enable users to conduct specialized queries of the locations of released prisoners, map those results at the address level and then overlay the results with additional spatially-enabled data sets, such as support services for former prisoners. The Urban Institute will oversee the evaluation, process documentation and dissemination activities. The two organizations have worked together for more than 10 years on numerous projects, including the Institute's Prisoner Reentry Mapping Network and National Neighborhood Indicator Partnership.
Using open-source software — MySQL, Ruby on Rails and Google Maps — The Providence Plan will build a tool that can be replicated by other communities with minimal startup costs and no need for ongoing geographic information system (GIS) technical support. As a Web-based application with a familiar architecture (Google Maps), the tool will have a more user-friendly interface and lower bandwidth requirements than typical GIS applications. These attributes — along with easy-to-use dropdown menus — will bring state-of-the-art mapping technology to nontraditional users operating at a variety of workstations (e.g., desktop computers, laptops and handhelds).
The Urban Institute's evaluation will include quantitative data from user Web logs and qualitative data obtained through user focus groups and one-on-one interviews. The evaluation will document the considerations that go into employing the application for community-supervision and service-provision purposes, which will help to inform the criminal justice field on the potential value and use of this spatial-analytic tool. Deliverables will include a final report, a technical manual to support replication by other communities and a research brief for publication in practitioner-oriented journals.
The Rhode Island Department of Corrections; the Rhode Island Department of Children, Youth, and Families; the Providence Police Department; and the Family Life Center of Rhode Island will be the initial users of the tool. Family Life is a nonprofit that helps ex-offenders and their families by providing long-term case management services starting before release from prison and extending for two months afterward.
The policy and practice implications for this project include a demonstrated improvement in the quality of parole and probation supervision through more efficient and effective caseload management, greater collaboration and problem solving between correctional and law enforcement agencies, and stronger and more intentional efforts to connect former prisoners to service providers that offer targeted support in the areas of housing, employment, substance abuse and other services.
