COPLINK Visualization for crime analysis.pdf

.COPLINK: Visualization for Crime Analysis Hsinchun Chen, Homa Atabakhsh, Tim Petersen*, Jenny Schroeder*, Ty Buetow, Luis Chaboya, Chris O’Toole, Michael Chau, Tom Cushna, Dan Casey*, Zan Huang University of Arizona, AI Lab {hchen, homa@} / *Tucson Police Dept. Abstract We have applied various visualization techniques to assist crime analysis. These techniques include a hierarchical list, a hyperbolic tree, a time line tool, a periodic pattern tool, a GIS tool and a self-organizing map. We have integrated some of these techniques into a single tool in order to display different and synchronized views of the same data. This tool is flexible and scalable to allow for the incorporation of other visualization techniques in the future. This paper describes the various visualization techniques we have applied in the law enforcement domain and that we are integrating into our ongoing COPLINK project. 1. Introduction Visualization techniques have been shown to be useful in various domains but have not been widely studied for applications in crime analysis where it is useful to identify the relationships among different entities such as people, vehicles, addresses and organizations. It is critical for crime analysts to retrieve, understand and analyze these relationships efficiently and effectively. There are some existing problems with visualization for crime analysis. Network charts are widely used by crime analysts but these are usually drawn manually. There are very few software applications available for drawing these networks and data exists in tabular format, which makes it difficult for crime analysts to see the big picture in order to draw conclusions easily. We have developed various visualization techniques to integrate into our ongoing COPLINK project (Chen et al., 2001). We describe each of these techniques and how we have incorporated some of them into one scalable visualization tool. We also explain some of the lessons learned when applying these techniques