In The Atlantic, Emily Badger looks at how the city of Memphis is fighting crime by doing data-mining. This brought to mind an Atlantic piece from four years ago by Hanna Rosin, which also looked at how numbers-crunching and crime-fighting in Memphis. The picture that emerged from that data was not something people were eager to see. Excerpt:

The inner city, where crime used to be concentrated, was now clean. But everywhere else looked much worse: arrests had skyrocketed along two corridors north and west of the central city (the bunny rabbit’s ears) and along one in the southeast (the tail). Hot spots had proliferated since the mid-1990s, and little islands of crime had sprung up where none had existed before, dotting the map all around the city.

Janikowski might not have managed to pinpoint the cause of this pattern if he hadn’t been married to Phyllis Betts, a housing expert at the University of Memphis. Betts and Janikowski have two dogs, three cats, and no kids; they both tend to bring their work home with them. Betts had been evaluating the impact of one of the city government’s most ambitious initiatives: the demolition of the city’s public-housing projects, as part of a nationwide experiment to free the poor from the destructive effects of concentrated poverty. Memphis demolished its first project in 1997. The city gave former residents federal “Section8” rent-subsidy vouchers and encouraged them to move out to new neighborhoods. Two more waves of demolition followed over the next nine years, dispersing tens of thousands of poor people into the wider metro community.

If police departments are usually stingy with their information, housing departments are even more so. Getting addresses of Section 8 holders is difficult, because the departments want to protect the residents’ privacy. Betts, however, helps the city track where the former residents of public housing have moved. Over time, she and Janikowski realized that they were doing their fieldwork in the same neighborhoods.

About six months ago, they decided to put a hunch to the test. Janikowski merged his computer map of crime patterns with Betts’s map of Section8 rentals. Where Janikowski saw a bunny rabbit, Betts saw a sideways horseshoe (“He has a better imagination,” she said). Otherwise, the match was near-perfect. On the merged map, dense violent-crime areas are shaded dark blue, and Section8 addresses are represented by little red dots. All of the dark-blue areas are covered in little red dots, like bursts of gunfire. The rest of the city has almost no dots.

Betts remembers her discomfort as she looked at the map. The couple had been musing about the connection for months, but they were amazed—and deflated—to see how perfectly the two data sets fit together. She knew right away that this would be a “hard thing to say or write.” Nobody in the antipoverty community and nobody in city leadership was going to welcome the news that the noble experiment that they’d been engaged in for the past decade had been bringing the city down, in ways they’d never expected. But the connection was too obvious to ignore, and Betts and Janikowski figured that the same thing must be happening all around the country. Eventually, they thought, they’d find other researchers who connected the dots the way they had, and then maybe they could get city leaders, and even national leaders, to listen.

Er, no. More:

 The “Gathering Storm” report that worried over an upcoming epidemic of violence was inspired by a call from the police chief of Louisville, Kentucky, who’d seen crime rising regionally and wondered what was going on. Simultaneously, the University of Louisville criminologist Geetha Suresh was tracking local patterns of violent crime. She had begun her work years before, going blind into the research: she had just arrived from India, had never heard of a housing project, had no idea which were the bad parts of town, and was clueless about the finer points of American racial sensitivities. In her research, Suresh noticed a recurring pattern, one that emerged first in the late 1990s, then again around 2002. A particularly violent neighborhood would suddenly go cold, and crime would heat up in several new neighborhoods. In each case, Suresh has now confirmed, the first hot spots were the neighborhoods around huge housing projects, and the later ones were places where people had moved when the projects were torn down. From that, she drew the obvious conclusion: “Crime is going along with them.” Except for being hand-drawn, Suresh’s map matching housing patterns with crime looks exactly like Janikowski and Betts’s.


If replacing housing projects with vouchers had achieved its main goal—infusing the poor with middle-class habits—then higher crime rates might be a price worth paying. But today, social scientists looking back on the whole grand experiment are apt to use words like baffling and disappointing.

Charles Murray? Bueller?