Steven Spielberg’s 2002 movie Minority Report gave us a glimpse into what a ‘pre-crime’ future could look like: a world where super-intelligent computers could predict the incidents that would occur before the criminals had even got out of bed. Twelve years on, the same idea is being employed in New York, though it’s burning buildings rather than murders that are the target.

The system works by pulling in huge amounts of data about each building in the city, including its age, its sprinkler network, its lift system and whether the property is occupied or vacant. By crunching through this data and looking at the trends of recent fires, the authorities can identify the areas that are at a high risk. Empty buildings are twice as likely to become ablaze, for example, while low-income areas are also more vulnerable.

Beating the flames

The technique is known as FireCast, and the New York City Fire Department’s director of analytics Jeff Chen and assistant commissioner Jeff Roth recently spoke about the innovation at the Big Data Innovation Summit in Las Vegas. Of New York’s one million+ buildings, around 3,000 suffer a serious fire every year. As only a third of the properties in the city can be inspected in person, it’s hoped that FireCast can make a significant dent in the number of emergencies – it enables the official inspectors to target high-risk buildings to visit, rather than choosing them at random.

In total, the prediction system weighs up 60 different factors that can contribute to a building being set ablaze. The basics of FireCast have been in place since May, but significant expansion is due during the first few months of 2014. “Ultimately, we should see the number of fires go down, and fires should become less severe,” Jeff Roth told the Wall Street Journal.

“This is going to make the city – and our members – safer than ever before,” promised New York fire commissioner Salvatore Cassano when the system was originally unveiled. “We want to arm our firefighters with the best information and technology available.” The Fire Department is partnering with IBM to produce proprietary handsets that can log and track pertinent information in real-time as inspectors and firefighters go about their work.

A bold move

FireCast is part of a broader initiative from the new Mayor’s Office of Data Analytics (MODA) set up in New York. The database of information collected by MODA covers many local authority services – the department has prepared plans to use 911 calls to track outbreaks of flu across the city, for example. Much of the information collected is available through the NYC OpenData website, including everything from restaurant inspection results to the locations of wi-fi hotspots across the city.

The New York Fire Department’s 2013 end-of-year report highlighted the improvements MODA had made to FireCast in partnership with veteran firefighters: “Working with the inspectors, MODA validated the statistical rigour of [the at-risk list], and also worked with a set of historical fires from the last several years to train the fire risk model and appropriately weight each of those inputs. Whereas the first version of the fire risk model had weighted inputs based on focus group discussion, the MODA model used a dataset of historical fires to hone the weighting of the risk inputs.”

Secret of success

Since its inception, MODA has achieved success in finding stores selling bootleg cigarettes and in efficiently removing trees flattened by Hurricane Sandy. Data about which restaurants had failed to register a carting service for removing cooking oil was used to bust those venues that were illegally dumping their grease into the city’s sewers. Meanwhile, MODA’s database is also being used to prioritise the 25,000 complaints of illegal conversions the authorities receive every year: information pulled from 19 different agencies is compiled to assess which of these complaints is likely to be of most interest.

“I think of us as the Get Stuff Done Folks,” the group’s Michael Flowers said in an interview with the New York Times. “All we do is take and process massive amounts of information and use it to do things more effectively. What’s impressed me most about this job is learning how insanely complicated this city is – that activity is reflected in the data and on an amazingly detailed level. What we’re really running here is an office of New Yorkology.”

Following the lead

Other major cities in the United States are following suit by tapping huge databases of information to cut costs and improve focus. The Problem Properties programme run in Boston uses a variety of data sources – including the volume of complaints, safety records, crime figures and tax data – to identify which properties should be visited by the police as a matter of urgency. By highlighting negligent landlords or difficult tenants, the police teams are able to distribute their resources more efficiently.

Then there’s the Virginia schoolboy Viraj Puri, who is hoping to develop a big data system to pinpoint bullying problems even if the victims feel afraid to speak out. Pulling in information from public social media updates, Puri has demonstrated a map that highlights which areas are experiencing the most severe problems. “Through using algorithms from social networking websites like Twitter, Facebook and Google+ we’ve been able to narrow down where bullying would occur most on the map,” he explained to the BBC.

As the technology involved continues to improve, both in terms of how quickly data can be reported and how effectively it can be applied, fire prevention could be just the start of a wave of big data schemes in our towns and cities. Arrests for murders that haven’t been committed are still some way off, but perhaps they’re getting closer all the time.