How New York City Used Big Data To Expose Illegal Housing

/**/ /**/

In 2002, when Michael Bloomberg took over as mayor of New York, the city faced a major problem with illegal housing. Housing ’conversions,’ or the altering of housing to fit more tenants, had become a burgeoning health and safety hazard.

’There are, of course, building codes and residential codes about how many wall plugs you have to have, how much open space you have to have, there were fire codes you have to care about,’ explained Timothy Persons of the U.S. Government Accountability Office. Persons used Bloomberg’s crusade against slum housing conditions as a case study in his talk on big data to attendees of the Innovation Enterprise Big Data Summit in Boston.

According to Persons, the major challenge Bloomberg faced was an immense amount of data from different city departments. To make matters worse, police officers used different codes from those used by city inspectors, who used different codes from the fire department's. There was a huge amount of data that needed to be normalized in order to be understood.

’And so by doing that, the group in New York City could look at these things and combine these 900,000 property lots and these data sets, and all these various things, and start to look for anomalies,’ Person said.

In the old ’framework,’ exposing housing units with temporary walls, illegal tenants, and other health and safety hazards relied on what Persons called ’gut checks.’ Essentially, a landlord or safety inspector would need to have his or her own suspicions. The success rate of gut checks was about 12 percent.

Following the installation of Bloomberg’s new data-dependent system, inspectors ’moved overnight from 12 to 72 percent,’ Persons said.

PUBLISHED SEPT. 14, 2015