Estimating Crime Rates Using Jumping Finite Automata on Tweets

In the Fourth Industrial Revolution, crime is hardly reported to the Police, or other law enforcement agencies. Most victims prefer to go to Social Media and vent, as this medium is easier for them to access and requires no paperwork or interrogations. This trend leaves policy makers and the law enforcers with skewed dataset, due to unreported crimes. Hence, it is paramount that one finds a way to “mine” the crime data reported on social media. In this paper, we have attempted to estimate crime rates, using Twitter as a data source. To do this, we have used a formal technique — Jumping Finite Automata (JFA), for the abstraction of a corpus of crimerelated words and used shuffle algorithms to establish semantic relationships between these words. The JFA was implemented in a tool called “Crime-Ripper”. Crime-Ripper uses tweets retrieved from crime hashtags on Twitter to estimate crime rates and produce reports that are map annotations, showing areas of a city and their respective estimated crime-rates. Crime- Ripper is expected to find applications in law enforcement, policy making and public safety sensitization.