Pattern Detection and Visualization for Strategic Crime Analysis

• In Event: Posters I

Abstract

Visualizing patterns of crime is challenging when considering the multitude of variables associated with spatial and temporal pattern detection. Current approaches leverage atomistic mapping solutions which fail to provide sufficient context beyond a limited perspective. This study takes a conceptual, holistic analytical approach to identifying both obvious and non-obvious crime correlates in urban locations with long-term problems. Using spatial statistical analyses methods and self-organizing maps (SOM), a variety of quantitative and qualitative attributes of urban actors were used to mete out observable and latent pattern discovery. A combination of both SOMs and traditional mapping techniques will be used to visualize the patterns for knowledge discovery.

Author

• Crystal English, San Diego State University