Data mining helped earn the Pulitzer Prize

The most prestigious journalism award this year was received by the Sun Sentinel newspaper from South Florida for its series “Beyond the Law: High Speed ​​Cops”.

For the investigation, the journalists requested from police stations records about the passage of patrol cars SunPass checkpoints on toll roads. Each SunPass entry is marked with the exact time of passage of the item. Police provided information without suspecting a trick.


SunPass checkpoint on toll road
')
The police did not think that if you measure the distance between neighboring SunPass checkpoints and knowing the travel time between them, you can calculate the speed of the car.

The reason for the investigation was the case when in the fall of 2011 one of the Florida police officers were stopped by traffic police for driving at a speed of 190 km / h. As it turned out, this is not an isolated case. Data mining of over a million records from SunPass checkpoints since 2004 revealed 793 police cars that were moving at a speed of 145-210 km / h, often during off-hours. The violators were the fifth part of the police fleet. More than half of the cases of speeding were registered outside the patrol zone, most of the vehicles were traveling along standard routes, most likely home after work or in the morning to work.

With the participation of police cars, at least 320 accidents have occurred since 2004, killing 19 people. But only one policeman was in prison and spent 60 days there.

The journalistic investigation caused a public outcry. The local police had to impose disciplinary measures and dismiss several employees. In some cities in Florida, police have introduced new systems to track employee actions during off-hours. Most importantly, after almost a year, the data mining showed a real change in the behavior of the police: they were less likely to exceed the speed.

The original Sun Sentinel data collection method was so effective that law enforcement agencies sent their experts to the editorial office with a request to share their experience. Perhaps they themselves will use such an algorithm to identify violators of traffic rules.