Even casual observers are well aware that Big Data is frequently used to create targeted ads, to recommend products for purchase by online shoppers, and to suggest connections on social networking sites. But people may not be aware that Big Data has become an important tool for law enforcement and organizations to stop crime before it happens. By connecting various datasets and analyzing them simultaneously, investigators are now able to identify patterns and norms, spot aberrations and red flags of criminal activity or conditions that might otherwise not have been detected using conventional methods, and to step in to prevent crime from occurring.
Police departments across the country are mining data from arrest records, social media and surveillance video to detect and predict crime patterns, which allows police to target their resources more effectively. The City of Los Angeles used data from 13 million crimes to predict where crime might occur on any given day. The LAPD credits its Big Data approach for a 33% decrease in burglaries, a 21% decrease in violent crime and a 12% decrease in property crime.
Police in London that are focused on “troubled families” analyze data from local authorities, social services, truancy reports and beat officers “to reveal the real trouble spots and troublemakers in a neighborhood, estate or street.” London police also have utilized advanced software to analyze 5 years of data to predict if and when a criminal will re-offend. The data includes geography, past offenses, and associations, and social media is monitored for inflammatory comments aimed at rivals or evidence of planning a crime.
The University of Pennsylvania Department of Criminology has developed an algorithm to analyze data and predict who is likely to be targeted for murder. They partner with police to warn the potential victims and provide advice on how to protect themselves.
Corporations are using Big Data and analytics to detect workplace situations that could lead to sexual harassment claims and to identify employees with incentives and opportunities to commit fraud.
Banking institutions utilize behavioral authentication, which tracks factors such as geographical location, how often a user accesses an account from a mobile device or computer, and how fast the username and password are entered, to prevent criminals from accessing a user’s account.
CNA uses open source Big Data tools and proprietary analytics to root out and prevent fraudulent insurance claims. Anomalous activity and patterns indicating fraud can be detected, and networked relationships can be uncovered before additional claims are submitted. They might identify that the claimant in one claim was previously the witness in a similar claim, which could be a signal that accidents are being staged. An individual claim that might not seem suspicious on its own may trigger additional investigation when it is examined in light of historical data.
As more and more institutions are expected to turn to Big Data analytics to prevent criminal activity, several concerns emerge:
Are steps taken to ensure that accurate data is being utilized in Big Data crime prevention efforts?
Are correlations being mistaken for causation? For example, when analyzing crime location data, is care taken to ensure that nearby residents aren’t mistakenly identified as the perpetrators, as opposed to the victims of crime, creating something of a cyber stop and frisk situation?
Who is developing the algorithms? Have the data scientists collaborated sufficiently with subject matter experts to ensure that the algorithm will serve the desired purpose?
Is the analysis neutral? Does the algorithm reflect any prejudices or biases of the researchers?
Are steps taken to ensure that individuals have not been forensically framed by a criminal who has obtained another person’s login credentials?
By taking care to ensure the integrity of the data, algorithms, and analyses, Big Data and analytics can be an effective and powerful weapon for law enforcement and corporations to target their limited resources more effectively and efficiently to predict and prevent crime.
This article originally appeared here. Republished with permission. Submit your copyright complaints here.