Sundeepta Achanta will defend her Master's Project, "Spatial Clustering Using Transportation Networks", on Wednesday, April 17 at 11:30 a.m. in 256C Avery Hall. Her advisor was Dr. Ashok Samal and committee members were Dr. Jitender Deogun and Dr. Leen-Kiat Soh.
Understanding the patterns of crime and their spatial distribution is important to develop management and prevention strategies. Identifying crime hotspots or clusters is therefore an important task. In this research, we propose a new distance measure that can generate clusters that are conceptually coherent. The distance, called the transport distance, is based on the path between two points in space using the underlying transportation network. This distance more accurately represent the conceptual distance that the traditional Euclidean distance. We have extended, DBSCAN, a popular density based clustering algorithm to LDBSCAN, which includes the transport distance. The distance measure and the clustering algorithms are implemented in ARCGIS environment and tested using a real-life crime dataset from the city of Lincoln.