Data mining to stop crime, or support it?
The New York Times had an interesting pair of articles yesterday about the use of data mining and its effect on criminal justice. The first one described the expansion of business intelligence software beyond its initial role in the retail world. Business intelligence software examines patterns in large amounts of data. For example, this allows Wal-Mart to optimize inventory in each store by taking into account local variations in taste.
In 2005, the police chief in Richmond, Virginia decided to try similar software to predict criminal activity. They analyzed 911 calls, arrests, and other data points. They found that crime increased in Hispanic neighborhoods on payday. The reason? Many residents do not use banking services, making the extra cash a tempting target for thieves. Using analyses like this, the police were able to better deploy their forces in advance of problems, and crime fell 20%.
The second story, entitled “Bilking the Elderly, With a Corporate Assist”, showed a darker side. Data mining has found a lucrative home in consumer marketing, allowing advertisers to target their pitches to the people most likely to buy. InfoUSA is an example of a […]