Cartel activities occur over space and time at one or more geographical locations, and are based on several possible functional building blocks. However, identifying the space-time characteristics (context) of such geographic locations and the relations between them is not simple, as such information can often be extracted only by exploring the convolution of activity type, actor, and locations over time. As a result, a complete and clear picture of the spatial footprints of cartel illicit activity cannot be obtained, limiting situational awareness and operations.
In order to address this challenge, CINA researchers collaborated with groups from our fellow COEs at the Center for Accelerating Operational Efficiency (CAOE) and the National Consortium for the Study of Terrorism and Responses to Terrorism (START) to form a multi-site team that pursues the derivation of actionable knowledge for cartel activities from open-source content.
The CINA portion of this effort is led by Arie Croitoru, Associate Professor, Geography and Geoinformation Science. Croitoru’s team is developing a linked illicit activity gazetteer and data analyzer. This gazetteer is constructed by building explicit links between place toponyms, possible colloquial synonyms, non-geospatial concepts and terms, and other (external) available open data and knowledge sources – in particular social media and news articles. The ensemble of these various links combines location information, relations to other locational information, event information, and non-geographic information into a knowledge graph. This graph can then be queried and mined to identify emerging patterns, or test user-driven queries. Summarizing the challenges behind this effort, Croitoru states that “The biggest challenge in our project is to discover the links between different locations which are often implicit. Addressing this challenge will enable delivering to intelligence analysts a much better insight into how cartels operate over space and time.”