VACCINE: Visual Analytics for Command Control Interoperability National Security and Emergencies

Project: Research project

Project Details


VACCINE: Visual Analytics for Command Control Interoperability National Security and Emergencies Developing a spatial statistics framework in the visual analytics for law enforcement technology suite This project will explore the integration of advanced spatial statistics for hypothesis testing into the VALET and iVALET framework. We will assess the integration of spatial scan statistics within the context of user interactions and explore combination of user selected regions with scan statistics for hypothesis testing. Given the runtime of spatial statistics, novel implementations and parallel structures will need to be explored for use on the iPhone and iPad. We will also explore the combination of density estimation with advanced statistical measures such as G* and LISA as a means of finding spatially significant regions of crime. These measures will be used for report generation and alert algorithms to direct end-users to interesting patterns within their categorical spatio-temporal data. Concurrently, we will also explore the application of other statistical measures on the temporal components of the data, setting up scripts for batch processing of categories stored in the database. In this manner, we will be able to facilitate an automated report generation which can be scheduled to push alerts to users iPhones (based on user privileges) which indicate events of interest. Finally, we will explore the aggregation and dissemination of this categorical data for use by the general public. In this task, we will explore comparisons to moving averages and predictive events to allow for an analogous crime weather map showing deviations from the normal expectations. The benefits of such work will be to enable analysts to statistically determine spatial anomalies (as opposed to purely visual exploration). The outcomes of this work will be the integration of more statistical analysis techniques into the VALET framework which will better enable analysts to detect trends, correlations and anomalies within the data. WDYTYA: Spatiotemporal Network Dynamics for Community Detection I. Overview Within social network data, there exist communities of users that grow and shrink over time. These communities share information and their network links and shared information can provide additional information about the spread of topics and information exchange in both the real and virtual world. These connections can provide direct clues as to the nature of an individuals identity and their role within both online and offline communities, allowing for the creation of cyber-geodemographic profiles. In order to extract such information, the Who Do You Think You Are? (WDYTYA) Spatiotemporal Network Dynamics for Community Detection (WDYTYA: SNDCD) project will explore the overall characterization of social network structure with respect to its relationship with geographic places. Key questions include the definition of new metrics for quantifying the influence and correlation of geographic locations on the formation of relationships and the degree of similarity of individuals and the quantitative analysis of these relationships over time. Such information can be directly used for outlier detection, as well as for profiling groups of people and/or entire urban areas or regions. With respect to Twitter, tweets can be generated using computers or mobile phones to which a geographic location might be associated. In particular, mobile devices are embedded in the physical world and are carried by people in their everyday lives. Therefore, these can be considered as links between physical and virtual worlds: in other words, it is possible to link information about physical and virtual users' behavior by analyzing their movement and their activity on online social networks and other websites. The characterization of the correlation and interdependency between virtual and physical networks will be an essential contribution of the proposed research work. Furthermore, location over time can be studied in terms of trajectories in a physical space. WDYTYA: SNDCD plans to quantify the similarity of users not only with respect to their locations at certain times but also with respect to their relationships in geographic space (locally and globally). The project team will define metrics for describing average characteristics of individuals and communities in Improving Predictive Analytics Capability for the Visual Analytics Law Enforcement Technology 1. Project with which this proposal is associated (type in below): a) Public Safety Please provide a brief abstract of your project (1-2 paragrpahs, 400 words maximum) In the past, we worked on increasing the analytical capabilities of the VALET software, adding in Morans I and Scan Statistics. This year I would like to explore the flow of criminal incidents over time, focusing on methods derived from volume rendering. Specifically, the goal would be to finalize research on displaying temporal trends and predicted temporal trends as flow maps. Simultaneously, we plan to continue work on the Top 10 List idea of crimes that officers should be policing for on a given day. We should be able to collect information on housing, census and businesses in the area and create similarity scores to weight the Top 10 list based on both a geographic regression and a sort of not-necessarily adjacent but still local similarity metric which relates crimes of similar neighborhoods to other neighborhoods to allow for improved policing. As part of this work, I plan to try to strengthen CVADA ties to the NYPD, currently discussions have begun for an MOU and future requests for modifying VALET to ingest their data. These partnerships were formed through connections with ASU and the GeoDa center here. Publications to be targeted should include not only visualization forums but also social science research forums as well. To this end it may also be interesting to document observed behavioral patterns from the given data as means of social science evaluation and focus on tool sets that can also enable this research simultaneously with predictive crime analytics.
Effective start/end date12/15/128/31/17


  • US Department of Homeland Security (DHS): $399,872.00


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.