Big Data Cop vs. Crime in Los Angeles

The fight against crime is increasingly achieved through present technological advances. Proof of this is the study and actual application of Big Data and Analytics, currently underway in California’s Los Angeles Police Department (LAPD).

Research Starting Point

To this end, they use a mathematical model originally developed by Professor Jorge Moher (Santa Clara University) to predict the aftershocks of earthquakes; just as when an earthquake takes place, there is a high probability that aftershocks will occur in a short time after the mainshock.

To test the hypothesis at police level, data on 13 million crimes committed in the city of Los Angeles in 80 years were analyzed. Whenever a crime had been committed, a higher likelihood that a new criminal offense would take place in the same area was observed. And they proved that the model pattern of those criminal activities was similar to the model used to predict earthquake aftershocks.

Model’s actual implementation

Through Big Data, the LAPD began in 2014 a pilot project for the implementation of Moher’s mathematical model; aiming at predicting the areas in Los Angeles where crimes would be most likely to occur daily. The algorithm identifies high-crime hot-spots in these areas, and the district police stations are notified of the likelihood of criminal offences and type, so as to prepare the police officers available for each shift.

The probabilities so obtained cover a maximum time span of 12 hours. Therefore, police patrolling is increased in the hot-spots within the geographical area where the likelihood of crime is higher during the prediction validity period. In order to measure performance and efficiency, the data is transferred in real time to the police headquarters.

Results

Through the Big Data software and Analytics, the police achieved a 33% reduction in the number of robberies, 21% in violent crimes and 12% in crimes against property, in the areas where the software was used.

The Los Angeles Police Department hopes that through continuous updating of data, the model’s predictions will become more accurate and will cover a longer time period, thereby allowing for higher accuracy in the detection and prevention of daily crimes in real time.

Original source: Datafloq

This entry was posted in News and tagged , , , . Bookmark the permalink.