Get SF Weekly Newsletters
Pin It

All Tomorrow's Crimes: The Future of Policing Looks a Lot Like Good Branding 

Wednesday, Oct 30 2013

Page 2 of 6

From the beginning, the effectiveness of PredPol had been a sticking point in the negotiations.

SFPD's Merritt was skeptical. In a series of e-mails from July 2012 to August 2013, Fowler laid out the technical specifications for the software and the types of crimes PredPol claims to predict. "The crimes we predict are burglary [residential, commercial, auto], auto theft, theft, robbery, assault, battery, and drug crime," Fowler wrote on July 23, 2012. "This goes significantly beyond your current ... mapping tools," he added. Affixed to Fowler's e-mail signature was the claim that PredPol's predictions are "twice as accurate as those made by vet cops."

Fowler's e-mails also make it clear that PredPol viewed the SFPD as a major potential contract that would drum up more business. The price, Fowler wrote on July 23, was "$150,000 with a 50 percent discount for signing up as one of the 15-20 early showcase cities nationally and a commitment for collaboration over the next three years." These showcase cities included Salinas, Seattle, and Alhambra.

Merritt pressed Fowler about whether the program could handle violent crime. "Homicide is a priority in the department — and if it is not there it would just beg the question why not," she wrote. Fowler admitted at the time that PredPol wasn't predicting homicides and gun violence. Failing to rope SFPD into its 15-20 city pilot program last year, PredPol devised a new idea, based on the demand for a tool to predict violent crimes, to include San Francisco in a gun-violence pilot program with Atlanta and Detroit. That would have used the three cities as case studies in its nationwide marketing push. When Detroit backed out, and San Francisco stalled, PredPol negotiated a deal with Seattle to cooperate on research into predicting gun violence. That collaboration was announced in May of 2013.

Following a favorable review of PredPol's pitch by Suhr, over the summer of 2012 PredPol negotiated access to SFPD's historic crime data and built PredPol's demonstration project for Mission Station, which mapped out locations where crime was most likely to occur based on historical crime data. Screenshots obtained from PredPol's internal server indicate the demonstration project was expanded to include all 10 of the city's patrol districts. However, Merritt says PredPol has yet to be used by patrol officers in the field.

From fall 2012 through spring 2013, Fowler repeatedly pressed Merritt for a set date for PredPol to be officially launched in San Francisco. Frustrated by SFPD's caution, and the department's more pressing goal of building its crime database warehouse, Fowler on May 9 lobbied Police Commissioner Loftus and Tony Winnicker, a senior advisor to Mayor Ed Lee.

Four days later, Merritt wrote to Suhr that the SFPD was not prepared to go public with PredPol: "Chief, I corresponded with PredPol and suggested that we not participate in this announcement at this time. ... While we will be rolling out PredPol and the gun violence module, I would like to wait until we are fully implemented before any announcements."

While the SFPD continues to weigh the merits of PredPol, more than 150 police departments nationally are deploying predictive policing analytics. Many departments are developing their own open-source algorithms, and a few tech heavyweights like IBM and Palantir are getting in on the game. But PredPol has emerged early to dominate the market. The company has sold its proprietary software here and abroad, from Kent County in England to Seattle, Wash., and here in the Bay Area to cities including Richmond, Los Gatos, Morgan Hill, and Santa Cruz. The origins of predictive policing, and of PredPol, however, are in Los Angeles, Santa Cruz, and Iraq.

The concept of predictive policing — forecasting where crimes are more likely to occur and attempting to prevent them — is rooted in the thinking of George Kelling, a theorist with the conservative Manhattan Institute. Kelling's ideas, published in several papers in the 1980s, centered around ways to better deploy limited police resources by using statistical analysis. The New York City Police Department's development and introduction of CompStat in the mid-1990s reorganized policing in the city to respond to trends in crime statistics, and to allocate officers accordingly. CompStat was a major change in American law enforcement. Police commanders increasingly made decisions based on data and statistical analysis rather than hunches. Crime rates dropped in New York following CompStat's introduction — although there is no consensus that CompStat was responsible. Nevertheless, the program became the gold standard for every police chief.

Virtually every police department in medium to large cities today has one or more crime analysts on staff to crunch numbers and plot past crimes on maps. Few had ever tried predicting future crimes though. Interest in predictive policing spiked nationally in 2009 as the National Institute of Justice, the research and policy branch of the Department of Justice, published a series of white papers and doled out millions in grant money to seven police departments to undertake the task.

One of the grants went to the Los Angeles Police Department in 2009. When LAPD applied for the grant two years earlier, it was still under the leadership of Bill Bratton, who had championed CompStat's introduction while serving as NYPD commissioner from 1994 to 1996. Bratton wanted LAPD to be a crime-fighting laboratory. He assigned then-Lt. Sean Malinowski, a former Fulbright scholar who had studied counterterrorism at the Egyptian National Police Academy in Cairo, to be the lead investigator on LAPD's predictive policing grant.

Around the same time, researchers at the Institute for Pure and Applied Mathematics at the University of California, Los Angeles were using grants from the Army, Air Force, and Navy to develop a series of algorithms based on earthquake prediction to forecast battlefield casualties and insurgent activities in overseas war zones. Army Research Office documents reveal that the work of anthropology professor Jeffery Brantingham, math professor Andrea Bertozzi, and math postdoc George Mohler was repurposed from its initial application of tracking insurgents and forecasting casualties in Iraq to analyze and predict urban crime patterns. This research would lead to the creation of PredPol.


Subscribe to this thread:

Add a comment

Popular Stories

  1. Most Popular Stories
  2. Stories You Missed


  • clipping at Brava Theater Sept. 11
    Sub Pop recording artists 'clipping.' brought their brand of noise-driven experimental hip hop to the closing night of 2016's San Francisco Electronic Music Fest this past Sunday. The packed Brava Theater hosted an initially seated crowd that ended the night jumping and dancing against the front of the stage. The trio performed a set focused on their recently released Sci-Fi Horror concept album, 'Splendor & Misery', then delved into their dancier and more aggressive back catalogue, and recent single 'Wriggle'. Opening performances included local experimental electronic duo 'Tujurikkuja' and computer music artist 'Madalyn Merkey.'"