Social media has found many uses, but one might hesitate to think that a message that is limited to 140 characters could be helpful in predicting crime. Sure, Twitter has made a major impact on our world… just ask anyone affected by the Arab Spring.
Not only has the social media phenomenon used to facilitate national revolutions, but analysts were able to predict successive uprising and predict the likelihood of success. Everything from the spread of infectious disease to anticipated receipts for a movie premiere have been calculated using the latter-day cabal called Twitter.
Enter Matthew Gerber, PhD, of the University of Virginia’s Predictive Technology Lab. His interest is in “…identifying structured information within unstructured text.” So, how did he do it? First, he started with existing crime statistics of neighborhoods in Chicago. He next organized tweets geo-located near the crime scenes and organized them by topics.
Constructing an algorithm, Dr. Gerber correlated certain topics or phrases to specific criminal activity (e.g. burglary, stalking, robbery, etc.). By using this catalog of words and phrases from Twitter, his model will predict how likely a certain type of crime is going to occur in another neighborhood.
This could lead to a significant change of Webster’s Dictionary…
- an insignificant, silly, or bothersome person.
- an inept criminal whose Twitter account got him in jail.
Bruce Bremer, MBA is LET’s technology contributor. Bruce retired from the Submarine Service after 21 years of in-depth experience with complex electronic technology. Since then, he has been involved in fleet modernization and military research analysis. He teaches electronics and alternative energy at a Virginia college. Besides his MBA, Bruce earned a Bachelor of Science degree in computer networking. He has been volunteering in public safety for many years.
Learn more about this article here: