TY - GEN
T1 - Visual analytics law enforcement toolkit
AU - Malik, Abish
AU - Maciejewski, Ross
AU - Collins, Timothy F.
AU - Ebert, David S.
PY - 2010
Y1 - 2010
N2 - We present VALET, a Visual Analytics Law Enforcement Toolkit for analyzing spatiotemporal law enforcement data. VALET provides users with a suite of analytical tools coupled with an interactive visual interface for data exploration and analysis. This system includes linked views and interactive displays that spatiotemporally model criminal, traffic and civil (CTC) incidents and allows officials to observe patterns and quickly identify regions with higher probabilities of activity. Our toolkit provides analysts with the ability to visualize different types of data sets (census data, daily weather reports, zoning tracts, prominent calendar dates, etc.) that provide an insight into correlations among CTC incidents and spatial demographics. In the spatial domain, we have implemented a kernel density estimation mapping technique that creates a color map of spatially distributed CTC events that allows analysts to quickly find and identify areas with unusually large activity levels. In the temporal domain, reports can be aggregated by day, week, month or year, allowing the analysts to visualize the CTC activities spatially over a period of time. Furthermore, we have incorporated temporal prediction algorithms to forecast future CTC incident levels within a 95% confidence interval. Such predictions aid law enforcement officials in understanding how hotspots may grow in the future in order to judiciously allocate resources and take preventive measures. Our system has been developed using actual law enforcement data and is currently being evaluated and refined by a consortium of law enforcement agencies.
AB - We present VALET, a Visual Analytics Law Enforcement Toolkit for analyzing spatiotemporal law enforcement data. VALET provides users with a suite of analytical tools coupled with an interactive visual interface for data exploration and analysis. This system includes linked views and interactive displays that spatiotemporally model criminal, traffic and civil (CTC) incidents and allows officials to observe patterns and quickly identify regions with higher probabilities of activity. Our toolkit provides analysts with the ability to visualize different types of data sets (census data, daily weather reports, zoning tracts, prominent calendar dates, etc.) that provide an insight into correlations among CTC incidents and spatial demographics. In the spatial domain, we have implemented a kernel density estimation mapping technique that creates a color map of spatially distributed CTC events that allows analysts to quickly find and identify areas with unusually large activity levels. In the temporal domain, reports can be aggregated by day, week, month or year, allowing the analysts to visualize the CTC activities spatially over a period of time. Furthermore, we have incorporated temporal prediction algorithms to forecast future CTC incident levels within a 95% confidence interval. Such predictions aid law enforcement officials in understanding how hotspots may grow in the future in order to judiciously allocate resources and take preventive measures. Our system has been developed using actual law enforcement data and is currently being evaluated and refined by a consortium of law enforcement agencies.
UR - http://www.scopus.com/inward/record.url?scp=78651507981&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=78651507981&partnerID=8YFLogxK
U2 - 10.1109/THS.2010.5655057
DO - 10.1109/THS.2010.5655057
M3 - Conference contribution
AN - SCOPUS:78651507981
SN - 9781424460472
T3 - 2010 IEEE International Conference on Technologies for Homeland Security, HST 2010
SP - 222
EP - 228
BT - 2010 IEEE International Conference on Technologies for Homeland Security, HST 2010
T2 - 2010 10th IEEE International Conference on Technologies for Homeland Security, HST 2010
Y2 - 8 November 2010 through 10 November 2010
ER -