Multi-Scale Analysis of Crime and Incident Patterns in Camden Dawn Williams {dawn.williams.10@ucl.ac.uk} Department of Civil, Environmental & Geomatic Engineering (CEGE), UCL Progressive Deepening Approach • Scale effects: the kinds of relationships or patterns that can be found are directly affected by the spatial or temporal granularity chosen 1. • A progressive deepening approach to the STDM 2 shall be applied to ESTDM • This will allow analysis through the continuum from micro to macro scales • This methodology can be applied to different datasets and processes by changing the tools and the spatial and temporal resolutions used. 1 Yao, Xiaobai. "Research Issues in Spatio-temporal Data Mining." Workshop on Geospatial Visualization and Knowledge Discovery. Virginia: University Consortium for Geographic Information Science (UCGIS) , 2003. 2 Han, Jiawei., 2003. "Mining Spatiotemporal Knowledge: Methodologies and Research Issues." Geospatial Visualization and Knowledge Discovery Workshop. University Consortium for Geographic Information Science, Virginia. Planar and Network Kernel Density Estimation • Kernel Density Estimation is a commonly applied interpolation and hotspot technique in spatial analysis and crime analysis 1. • A Planar Kernel Density estimate was conducted using a bandwidth of 310m and a cell size of 40m. • Both the planar and network KDE were useful in identifying the clusters though the network KDE algorithm appeared to be less dependent on input parameters • Planar and Network KDE revealed hotspots in the vicinity of Holborn and Camden Town. Hagenauer, J., Helbich, M., & Leitner, M., 2011. Visualization of Crime Trajectories with Self-Organizing Maps: A Case Study on Evaluating the Impact of Hurricanes on Spatio-Temporal Crime Hotspots. In: International Cartographic Conference ICC 2011, International Cartographic Association (ICA), Paris. 1 Spatial Scan Statistics Overlapping cylinders are obtained. A crime hotspot or cluster is a cylinder in which the number of observed crimes is significantly larger, statistically, than the expected value. Clusters were labelled statistically significant when p ≤0.01 for both methods. The space time permutation algorithm found no significant clusters for the entire period when tested against a significance level of 1% i.e. p=0.05. Spatial Scan Statistics • The space-time permutation and space-time Poisson models were used. • The ST Poisson model found two clusters as expected. Both clusters had duration of 14 days but the start and end dates differed. • The Camden area cluster (red) had a smaller radius and a greater relative risk, than the wider cluster in the Holborn, Bloomsbury(orange) area. • This visualization technique is extremely effective and useful with the results being clear and easy to interpret. The Self Organizing Map Method • The Self Organizing Map (SOM) is an unsupervised, learning neural network1. • It allows simple, geometric relationships to be produced from vector quantization analysis of complex multidimensional datasets2 through a combination of clustering and dimension reduction3 while preserving topology2. Reorderable Matrix U Distance Matrix Map Matrix Parallel Coordinate Plot 1 Hagenauer, J., Helbich, M., & Leitner, M., 2011. Visualization of Crime Trajectories with Self-Organizing Maps: A Case Study on Evaluating the Impact of Hurricanes on Spatio-Temporal Crime Hotspots. In: International Cartographic Conference ICC 2011, International Cartographic Association (ICA), Paris. 2 Kohonen, T., Hynninen, J., Kangas, J., & Laaksonen, J., 1996. SOM_PAK: The Self-Organizing Map Program Package. Helsinki University of Technology, Otaniemi. 3 Guo, D., Chen, J., MacEachren, A. M., & Liao, K., 2006. A Visual Inquiry System for Space-Time and Multivariate Patterns (VIS-STAMP). IEEE Transactions on Visualization and Computer Graphics , 12 (6), pp. 1461-1474.