The Detroit Crime Commission (DCC) and Chenope are seeking several students to participate in a first-of-its-kind ongoing initiative to automatically categorize social media posts that are associated with violent criminal intent. This is a very interesting and difficult problem due to the diversity of language used to express criminal activity, the fact that the criminal actors won't always be explicit about their actual intent, and the fragmentary nature of social media. The expectation is that the initiative will spread to include other law enforcement agencies; the DCC is quite active in adopting advanced technologies to help fight crime, and promoting these technologies to other law enforcement agencies. This is a highly unusual opportunity to both help a worthy endeavor and to be involved in really interesting work.
The categorization requires capturing language usage in real world data so as to differentiate references to criminal behavior from harmless commonplace things such as jokes, sarcasm, references to TV shows, etc. The results produced by the software must then be compared to the evaluations of a knowledgeable human analyst so as to ascertain their accuracy. This is a considerable amount of work, which is why we are seeking to broaden the available resources to complete the work. However, for students with the right background there may also be an opportunity to participate in related research, such as how social network analysis (SNA) done in conjunction with the automated categorization can help further predict criminal incidents.
While the effort was begun last year, in this year it is expanding to include Countering Violent Extremism (CVE.) CVE relates to identifying terrorist messaging including recruitment attempts. Chenope expects to put the first dedicated CVE monitoring system in place in the U.S. in the Metro Detroit area; the timing will depend on available funding for the effort. Chenope has also started to work with some European CVE entities and the scope of the project may expand accordingly.
Participating students will be asked to sign an agreement to maintain the confidentiality of the Pii they will be exposed to during the project, as well as any techniques used by DCC that they wish to remain private. Within these understandable limitations, DCC encourages publication of their work, as publicity helps attract both the participation of other law enforcement agencies and donors; DCC is a 501c3, and thus relies on charitable contributions.
Salary would be $12-14/hour for graduate students.
Internship Logistics & Parameters
The internship requires a few hours per week per student - exactly how many will be determined by the number of volunteers who sign up and how productive they are. Participation will not require travel to Detroit. DCC will provide the law enforcement domain knowledge. Chenope will provide the technology and CVE knowledge and oversight remotely from California and Edinburgh, Scotland.
The Detroit Crime Commission conducts research, assists the Detroit area police departments in investigations, disseminates information to the public, and helps coordinate crime reduction activities between business, the public, government and law enforcement. For more information, check out: www.detroitcrimecommission.com .
Chenope is an employee-owned software company founded by U of M Honors Math alums which specializes in analyzing large scale data for investigative purposes. For more information, check out: www.chenope.com.