The 10th Crime Mapping Research Conference was held August 19 – 22, 2009 in New Orleans. The CMRC is organized every one to two years by the Mapping and Analysis for Public Safety (MAPS) office in the National Institute of Justice. The CMRC brings together academic researchers, crime analysts and command staff to review the state of the art in geographic analysis and visualization of crime. This was a lively group with a number of high quality talks and workshops, and a strong series of research-oriented presentations. While sponsored by the NIJ in the U.S. Dept. of Justice, attendees included folks from Japan, Turkey, Canada, and the UK.
While there were several interesting presentations, I would like to highlight four that I particularly enjoyed:
I attended two presentations by Dr. Elizabeth Groff, a Professor of Criminal Justice at Temple University. In her keynote address, she suggested that while our ability to visualize crime patterns has steadily improved, there is a difference between information and “actionable” information, and, as a community, we need to be creating more tools that are aimed at generating actionable intelligence. She set out several examples of what this might look like in different operational contexts. I was very proud that one of her examples was Azavea’s prototype for HunchLab, the Crime Spike Detector currently in use at the Philadelphia Police Department. Dr. Groff also did a fascinating presentation introducing the use of Agent-based Modeling and how this simulation technique can be applied to modeling geographic patterns of criminal behavior.
Dr. Wilpen Gorr, from Carnegie Mellon University, presented a paper on Receiver Operating Characteristics (ROC) for Hot Spot Analysis. ROC was a technique originally developed for analyzing radar signals and has since been used in medical imaging, but crime analysis is a novel application. He has been working with the technique to calibrate kernel density smoothing as well as to use leading indicators to create predictive analytics for particular classes of crime.
Jerry Ratcliffe, a colleague of Liz Goff at Temple University, also did multiple presentations. The one I attended was on a pattern his team has been researching, called the “near repeat” effect. He has demonstrated that for some crimes, like burglary, there is a “contagion” effect that raises the likelihood of a repeat crime occurring near the original crime within a short period of time after the event. His work does not end with a research paper – he has built some helpful software tools to help calculate the extent of the effect.
In summary, this was a terrific event, and I have to commend the Director of MAPS, Ron Wilson, and his colleagues for creating an excellent forum for exchanging ideas.
