In his 1956 novella The Minority Report, author Philip K. Dick envisions a society in which telepathic mutants foresee crime before it occurs.
In studying the US city of Chicago, Illinois, which has what the New York Times calls a “notorious crime problem”, researchers are using more prosaic means – they are analysing statistics to find ways of accurately predicting the occurrence of crimes ranging from aggravated assault and battery to burglary and motor vehicle theft.
Their analysis, published in the journal PLOS ONE, revealed a surprising pattern.
“Most types of crime in Chicago had very distinct patterns by time of year, time of day, day of the week and even holidays,” says lead author Sherry Towers, from Arizona State University.
The research indicated that aggravated assaults and batteries occurring outside were more sensitive to climate variations than similar crimes occurring in homes, regardless of the time of year or the time of day.
“Days that are warmer than average significantly increase aggravated assaults and batteries in both locations, but the effect is more pronounced in street locations,” the report states.
The scientists say the trends thus evident in the recent past provide clues to reducing crime in the future. Taking into account temperature predictions, for example, may “significantly improve short-term crime forecasts for the temporal trends in many types of crime, particularly aggressive crime”.
The authors found a strong correlation between aggressive crimes and the hot weather of June and July during the northern hemisphere summer. “The confluence of hot summer days and weekends is thus a perfect storm that results in spates of shootings,” Towers says.
“If you can tell police where and when to focus their resources, it empowers them with information to be more effective and reduce crime,” adds research team member David Ebert, from Purdue University, in Illinois.
“There are natural time patterns to human activity, but they vary at the neighbourhood level so it is important to pick those out and equip officers with this information,” he adds. “We just don’t want to put the cops on the dots but put the right officers in the right location at the right times to be on the lookout for certain activities.”
The researchers examined 5.7 million reported crimes of violence or property-related incidents that occurred in the city of Chicago between 2001 and 2014. Towers says they chose Chicago as the target because it is “one of the first cities to make their crime data available to the general public”.
“They have the longest time series available, going back to 2001, and the largest number of crimes recorded,” she says.
“Their database is also very detailed, including details on the time of each incident, and what kind of location it occurred in – the residence, street, etcetera. The data were thus well suited for our study.”
Towers notes that although news reports concerning Chicago crime focus on shootings, the per capita rate of such incidents is actually lower than in several other American cities.
“Chicago isn’t even in the top five,” she says. “Cities like St Louis, Baltimore and New Orleans have much higher rates. But, Chicago has a much larger population, which means that even though its rates are lower than those other cities, the sheer numbers are higher, and that body count is what catches people’s attention.
“Moreover, patterns based on weather and time of day are important in all communities – even small towns.”
The study also took into account factors such as public and school holidays, days of the week, and pay days.
The researchers believe theirs is the first analysis of its kind to examine the potential effect of climate and holidays from a predictive analytics perspective. The climate variables they examine include temperature, humidity, wind speed, air pressure and precipitation.
They also found rainy and windy days tended to suppress crime. “People simply don’t want to go out in bad weather,” Towers says.