by CARLY BERWICK
On a tree-lined street on the west side of Jersey City, a man sat in a small sedan and shoveled white powder into small bags. Parked behind the car was a white sports car with its lights on. Every few minutes, the man would get out and bring several small baggies to the sportscar.
A woman watching from a window above called Jersey City police dispatch to report a drug deal at 2:51 am. The dispatcher asked for the address. The caller provided the street and the nearest cross street, uncomfortable with sharing her home address.
According to Jersey City police dispatch records for that particular date in November 2014, a car was sent to the cross street at 2:56 a.m. in response to a “suspicious condition.” Police dispatch records in Jersey City are voluminous, logging 169,482 total calls in 2014, a number that has increased each subsequent year. Police track the calls in a database and share them with the city’s information team, which puts the spreadsheet on the city’s Open Data Portal, which is free and available to anyone with internet access. The number of actual calls for service, however, is somewhat lower, once administrative uses of police cars, duplicative cars sent to the same instance, and error calls are factored in, making it more like 100,000 to 140,000 annually or 11-16 community calls for police an hour on any given day. As one of the most common data sets on municipal open data platforms, calls for police service exemplify the tension between openness and privacy that cities must navigate as they push more data online.
Calls for police service exemplify the tension between openness and privacy that cities must navigate as they push more data online
Source: thelocaljc.org; data from Jersey City Open Data Portal
Jersey City’s publication of its police calls for service on the Open Data Portal is part of a decade-long national trend toward open municipal data. Cities around the country are investing in “data dashboards” and platforms. Large cities such as New York, Los Angeles, and Chicago have well-staffed departments devoted to filtering, hosting, and maintaining data and to setting standards and protocols surrounding data use, following mayoral executive orders that all public city data be accessible online. A 2014 McKinsey report claimed nearly $3 trillion in global benefits to government “liquid information,” open, accessible, and “machine-readable,” at no cost to the public.
“When you are thinking about operational efficiency and workflow the calls for service is incredibly valuable,” says Brian Platt, Jersey City’s former Chief Innovation Officer and current chief business administrator. When he started five years ago, Platt looked at where officers were stationed relative to calls. “If there are 25% more calls in one area, we asked, why don’t we allocate more officers in that area?”
Calls-for-service in New York City (8.6 million population) are relatively even more voluminous (6 million in 2017), well-maintained, and updated quarterly. “Having access to the data through open portal is a simpler way to get the information you need to measure crime in neighborhoods,” says Sheyla Delgado, deputy director of the John Jay Research and Evaluation Center at the City University of New York. “The accessibility is great, and you can create indices.”
Source: Jersey City Open Data Portal, 2018
Particularly useful, says Delgado, is the location data. New York City publishes the latitude and longitude of each report but not the exact address of each call; Jersey City publishes both. While latitude and longitude can show building-level detail, the address is not searchable in the data table. It does offer a block-level map of where each report occurred. “That allows you to aggregate those points, so we do analysis at the neighborhood level, the census tract level, the borough level. We can create monthly counts of crimes for each census block in New York City,” says Delgado, who is conducting a large study of which interventions help prevent crime in public housing. The geographic information is critical, as officers often inadvertently record incorrect housing complexes for recorded crimes--up to 40 percent of the time in New York, according to Delgado. The only way to know if the intervention program works is to know if crime goes down in the housing complexes receiving the interventions. Without the location data, the study would be worthless--and researchers and police would have one less tool available to prevent crime.
Source: NYC Open Data, 2018
Privacy and profiling
Yet as open data increasingly becomes critical to informed and engaged citizen action and transparent and efficient government, some researchers and advocates worry that individual privacy may be lost in the rush to make data public. A 2017 Harvard Berkman Klein Center report identified this central tension between data and privacy: the best data is the most detailed (or “granular”), but this is also the data that can most put personal information at risk. Jersey City leaves out highly sensitive calls, related to child or domestic abuse, for instance, yet leaves in street addresses. Fear of having one’s address recorded and made public could potentially depress reporting. One woman posted this May on the Lincoln Park Neighborhood Facebook group complaining of a group of young men drinking, smoking, and gambling on her corner: “I call police and they wanted my name, phone. They did come but the group hide . . . I will call police [only] if somebody getting hurt.” Neighbors responded that she was not obligated to provide her phone number--something she hadn’t known.
Increasingly, open data efforts encounter concerns about digital equity--that is, making sure data sets aren’t used to overpolice or over-penalize poor communities and communities of color, says Greta Byrum, co-director of the Digital Equity Laboratory at The New School. Open data could potentially be misused to target surveillance by law enforcement or government, says Byrum, who points to Minority Report-style “predictive policing” algorithms used by departments in New Orleans, Los Angeles, and elsewhere and run by Palantir, which also has a contract with ICE to find potential undocumented immigrants (with an uncanny Jersey City/JP Morgan/Secret Service reference, in this Bloomberg report).
Crime reports, including Uniform Crime Reporting statistics, inform local investment decisions, with real-estate sites such as Trulia offering risk scores by neighborhood, citing JCPD service calls. “You could see real estate losing value based on how things are aggregated in publicly available algorithms,” Byrum adds. “Openness was accepted as a common good for a while, but now things are changing. Folks who are really worried about privacy, especially for at-risk and vulnerable communities, are less accepting of openness as a default public good.” Crime is reported in the southern parts of the city--particularly in Ward F and the Jackson Hill neighborhood--at at least twice the rate of the rest of the city. These neighborhoods are lower income, according to the Census, and less likely to have residents who have graduated high school than other areas in the city, particularly downtown or the gated communities of Port Liberte and Society Hill. Neighborhoods with more crime reports more may also be more vulnerable to real-estate speculation and underbidding; these areas were also the focus, for instance, of the no-knock legislation and campaign last year, which made it illegal for real-estate speculators to ask residents to sell their homes once they sign up to be on a no-knock registry.
Source: US Census Bureau, Census Explorer, 2018
As solutions, digital equity advocates ask for security audits, dumping stores of personal data, working with communities to inform them about open data and broadband access, and developing consent policies around data collection points. In Detroit, where the mayor’s office launched an open data initiative in 2015, community activists have been working to make sure open data does not lead to more surveillance and criminalization of low-income communities of color. The Detroit Digital Justice Coalition offers a guide for equitable open data. “While open data positively boosts government transparency , it can also be a source of fear and harm to residents,” the guide says. “We heard deep-rooted concerns about how the Improve Detroit Open Issues queue can incite property swindles, how Blight Tickets can reinforce ‘bad’ and ‘undesirable’ neighborhood reputations and influence property values or redlining practices, or how randomized block-level location of a Crime Incident isn’t random enough . . . to truly protect from re-victimization, stigmatization or further harassment.”
For mid-to-small sized cities such as Jersey City, the cost and human labor to power data platforms can be difficult to maintain, without the benefits of scale. In Jersey City, the data available includes numbers of people moving in and out of public housing, parking tickets issued, vacant buildings, and meetings of the Office of Emergency Management. Many of these resources, however, are posted by month or separate from visualizations, making it difficult for all but the most dedicated searchers to make sense of them. Some agencies providing information for the data platform don’t yet have fully digitized information, as well. “The issue with us and many cities is that the system isn’t typically as electronic as we want,” says Platt, the city’s former Chief Innovation Officer.
The dispatch data is, which makes it some of the most comprehensive and searchable. “Police data is good data, so everyone digs into that.” Platt mentions that a few years ago companies used to try to price municipal data dashboard platforms in the hundreds of thousands; now they are more like $30,000. “Cities of this size are in this weird gray area where they are not at that level of digging into data,” of the larger cities, he adds. “At end of day, [getting accurate information] is not a police problem but a data problem.”
One of the more crucial data sets for full transparency may be municipal budgets and salaries. Many larger cities’ platforms also allow users to see these details; Jersey City’s platform currently shows an overview but not specific salaries. New York City’s Department of Information, Technology, and Telecommunications has 1750 employees and a $635 million budget--larger than that for the entire city of Jersey City, which is $588 million. In Jersey City, the department of innovation started with a $2.25 million Bloomberg Philanthropies grant in 2014 to last for three years. That money ran out this year, and Platt became the city’s chief operating officer/business administrator, where he continues to work with the seven-person i-team. Funding for that comes out of the department of administration, whose proposed budget for 2018 is $2 million for the office of administration and $1.9 million for information technology.
Seattle (IT budget $36 million; overall budget $1.1 billion; population 700,000) is one large city--Chicago and New York are others--that pivoted to focus more on privacy when making data public, through establishing a Data Privacy Officer and an "open by preference" policy. At a municipal conference in December, Seattle's chief technology officer told attendees, according to Government Technology magazine, “Our people who are paying attention would much rather see us err on the side of not releasing something that could cause a mosaic effect or otherwise cause harm." To test the impact of data release, the city posts a limited set, observes and solicits public response, then releases more. This approach takes more people, who cost more money, however. Seattle's population is 2.7 times that of Jersey City, but its IT/innovation budget is at least 9 times bigger. In smaller cities that want to do right by both information and people, the money question will continue to vex planners and technologists. In the meantime, they keep uploading, cautiously.
Coda: Conversations and street signs
In the end, data is only as accurate as the humans recording it and monitoring it. In response to the drug deal beneath her window, the woman called dispatch three times over thirty minutes. According to the calls for service data, one car was sent to check on “suspicious conditions.” The caller--this writer--never saw a marked car.
Some neighborhoods have more crime than others, but crimes also have a relationship to the organization of space in a neighborhood. At the time of the drug packaging in front of my house, there was an unused reserved handicapped parking spot that had been installed for a neighbor’s tenant a few months earlier, who also had a driveway spot. The open space had become convenient parking for anyone looking for a guaranteed temporary spot in the middle of the night.
As Jane Jacobs wrote in her 1961 classic The Death and Life of Great American Cities, haphazard interventions can determine the life of a street. The unused handicapped space reserves a portion of a public street not for the person who may have originally needed it but for other spot sitters. In doing so, it turns away visitors, residents, and college students who had been eyes on the street. “The safety of the streets works best, most casually, and with least frequent taint of hostility or suspicion precisely where people are using and most enjoying the city streets voluntarily and are least conscious, normally, that they are policing,” wrote Jacobs.
When I saw the sedan parked there again, I called the West District precinct instead of dispatch to request an undercover car, which arrived promptly, trailed the car, and pulled him over. This time, it turns out, the driver had been using our side street for romantic purposes. His late-night parking spot disappeared soon after. The neighbor’s tenant moved, and the city removed the handicapped parking signs, leaving only the two metal poles as a record to the year the space below my windows served as a drug dispensary. None of those interactions were collected as public data.
Photo by Charles Deluvio on Unsplash