minority_report SpinOff

Scientists in London are claiming that mobile phone data can predict future crime hotspots with 70% accuracy. The findings provide evidence that aggregated and anonymized data collected by a carrier’s mobile infrastructure can contain relevant information to describe a geographical area in order to predict its crime level.

According to work done in Italy, Spain and at the Massachusetts Institute of Technology (MIT), daily data from mobile phones significantly improved the accuracy of crime predictions.

Unlike current systems that use crime statistics and local demographics which can be difficult and expensive to gather and not regularly updated, mobile phones can collect data about their owners gender, age and location of the phone in real time.

The proposed approach could have clear practical implications by informing police departments and city governments on how and where to invest their efforts and on how to react to criminal events with quicker response times. From a proactive perspective, the ability to predict the safety of a geographical area may provide information on explanatory variables that can be used to identify underlying causes of these crime occurrence areas and hence enable officers to intervene in very narrowly defined geographic areas. – “Once Upon a Crime: Towards Crime Prediction from Demographics and Mobile Data”

The scientists used data obtained from Telefonica, a European mobile phone company which owns the O2 service in the UK, to piece together an algorithm to predict crime rates in the next period of time. Included in this algorithm was information from the London Borough Profiles Dataset showing an area’s housing market, political affiliation, transportation, homelessness, life expectancy and other factors.

The study of the impact on behavioral development of factors like exposure to specific peer networks, neighborhood characteristics (e.g. presence/absence of recreational/educational facilities) and poverty indexes, has provided a wealth of knowledge from both individual and collective standpoints. Existing works in the fields of criminology, sociology, psychology and economics tend to mainly explore relationships between criminal activity and socio-economic variables such as education, ethnicity, income level, and unemployment.” – “Once Upon a Crime: Towards Crime Prediction from Demographics and Mobile Data” 

The author’s came up with predicted crime maps of London, UK:

LondonCrimeMap Source url LondonCrimeMap2 Source url

A negative aspect of this study showed how “anonymous” data really isn’t “anonymous” as information given to these scientists was used to track specific individuals. The scientists also admit to needing significantly more data and time to refine such a system for public use.

The methods found in this study really aren’t that much different from from those found in the movie, The Minority Report. Therefore, expect Tom Cruise to come and grab you in the middle of the night if you plan on committing a crime.