The DA’s office wants to use predictive analytics software to direct city resources to ‘places that drive crime.’ Will it work?
20th November 2023 · 0 Comments
By Nick Chrastil
The Lens
Lower 9th Ward native Brandi Charlot took a family-owned shotgun house and turned it into her shop, Blucid Floral. The little aqua-painted structure on St. Claude Avenue, between Forstall and Lizardi Streets, is a sought-after gem for flower lovers.
The blocks around it need some work.
As Charlot walks down her front step, her view to the right is two longtime vacant lots leading to a busy bus stop on the corner. Across the street is a beige strip mall whose tenants include a washateria and Brothers, a 24-hour gas station. Next door is a shuttered bar room, the once-popular Mickie Bee’s Lounge. A few blocks toward the river is another bar, Mercedes Place, and a restaurant, Cafe Dauphine.
All three have, or had, video poker licenses.
On nearby blocks, lines of tidy homes alternate with blight: vacant lots, empty buildings tagged with graffiti, and homes that have fallen into disrepair.
With its new predictive analytics software, the Orleans Parish District Attorney’s office can plug in the location of the bus stops, bars, restaurants, a wide range of businesses, video-poker machines, and other factors suspected to have a correlation to crime.
In response, the computer spits out a map of several areas at “high risk” for weapons crime.
Across the street from Charlot’s shop, the gas station is highlighted bright red, with shades of orange surrounding it. Nowhere else in Lower 9 is quite as risky, though two other gas stations blocks away also show up as potential problem areas. Locations deemed “environmental risk factors,” are denoted by symbols on the map – a gas pump, beer mugs, a restaurant place setting, and slot-machine cherries.
Farther toward Fats Domino Avenue, a stretch of St. Claude is dotted with geographical indicators of a school, two tire shops and a meat market. It’s highlighted in green, indicating low-risk for those specific weapons crimes.
The theory behind Risk Terrain Modeling (RTM) is that environmental factors have an outsized effect on crime in certain areas – and that even modest, strategic changes can deter future illegal activity. With the help of a federal grant, the DA’s office purchased a three-year, $127,000 subscription to RTM software. The project’s big-picture implementation is spearheaded by Andre Gaudin, a long-time New Orleans prosecutor who serves as the chief of screening under Jason Williams.
Yet the program’s skeptics question whether Gaudin and the DA’s office can mobilize a city government known for being unresponsive to citizens. Even critics who believe that policing here could use an infusion of new energy and data-proven strategies worry that the RTM maps – and its theme “places not people” – could give law enforcement a racially-neutral cover for singling out residents of resource-poor neighborhoods, which more commonly includes more “negative” factors, such as video-poker machines, tire shops and barrooms.
So far, though, the program’s hype and criticism are all hypothetical.
Last month, the DA’s office convened the first meeting of the program focusing on three New Orleans Police Department districts – the First, Fifth and Eighth – covering most of downtown New Orleans.
If the pilot is successful, the program will expand to the entire city.
***
The safest communities are communities have the most resources’
At first glance, some insights of the RTM software may seem obvious: for instance, everyone in the Lower 9th Ward already knows those blocks of the neighborhood struggle with gun violence.
But officials with the DA’s office have big hopes for the software. They believe that, by detecting correlations between specific geographical features and crime, they can identify places where crime is already bad, even places where crime happens but is not reported. Then they can go two steps further, to predict where crime might move in the future and how to preemptively intervene to prevent it.
The predictions are possible because people involved with crime often migrate to places with similarly risky geographic features. “We know that the conditions for the same aspect of crime are just four blocks away, and another spot 10 blocks away,” Gaudin explained. “So unless we fix all three of those with the same strategy, we’re just going to whack a mole from location to location.”
But the mole is unlikely to be whacked to high-income census tracts.
Or so says Sarah Omojola, director of the New Orleans office of the Vera Institute of Justice. Earlier this year, Omojola attended a Risk Terrain Modeling workshop at the DA’s office and she has participated in discussions about the upcoming pilot.
But Omojola is less convinced of the software’s predictive power. “We know that the safest communities are communities that have the most resources,” she said. “What Risk Terrain Modeling would do is show you where we don’t have resources.”
The DA is able to subscribe to the RTMDx software and pay for several data-analyst positions thanks to a $1.3 million U.S. Department of Justice grant issued to the office last year. The grant is part of a program to combat drug abuse and overdose deaths — and to ostensibly increase participation in drug abuse diversion programs. But to start, OD-prevention efforts require data-sharing agreements. As a result, the software will get its debut with crime-risk mapping.
To get the right results, you need the right mix of data, critics say. The DA’s office already has access to a wide range of anonymized NOPD data – including records of field interview cards that officers fill out when they interact with the public, NOPD calls for service, police reports, and arrest records, according to an office memo. In the future, the program will also use EMS and fire department calls for service.
The DA’s list of potential risk factors is also broad: it includes dozens of types of businesses and public facilities, from barbershops to grocery stores, bus stops and schools. In addition, the office says they can draw correlations between certain crimes and 311 calls for service related to infrastructure and quality-of-life concerns like potholes, abandoned vehicles, broken street and traffic lights, and noise complaints. Project planners also hope to include data from the City’s Code Enforcement that would show blighted properties and the prevalence of short-term rentals.
Police district commanders and neighborhood leaders would direct the initial RTM analysis. For example, they might ask for a closer look at geographic risk factors tied to carjackings in a certain neighborhood, or at a certain time.
The office would run the requested data and bring it to the group meetings, to discuss why certain locations might be risky at certain times and to determine ways to make them less so. Fairly simple interventions could include fixing a blighted property or moving the location of a bus stop.
“We don’t have to spend our resources fixing every streetlight,” Gaudin said. “But if we fix the streetlights that are near those locations, right, we’re gonna maximize our bang for our buck .. and we might actually see a decrease in the amount of arrests for that particular crime that we’re targeting.”
Policing could also shift, through patrols focused on particular blocks, regular officer check-ins at businesses, or use of the city’s recently passed “padlock ordinance”, with the help of the city attorney’s office, to shutter properties the police deem “chronic nuisances.” (The office recently demonstrated another tactic for those problem businesses when it used asset-forfeiture laws to seize an Iberville-neighborhood corner store deemed a “hotspot.”)
Ultimately, at a city level, RTM could guide priorities, resource allocation, and budgetary decision-making, making public safety what Gaudin calls “the lodestar for the entire apparatus of city government.”
***
The need to rebuild trust and social fabric, along with infrastructure
Charlot’s family has occupied her flower shop’s building for several decades. So when she speaks about the Lower 9 prior to Hurricane Katrina, her memories are about the spirit of the place. She remembers vibrant businesses and neighbors who knew and trusted each other.
“We played outside,” Charlot said. “We mingled with friends. We had neighborhood police. I mean, you could go next door and ask somebody for bread. Bread and sugar. You can’t do that now.”
In 2005, the Industrial Canal’s levees broke as Hurricane Katrina approached, flooding much of the neighborhood up to rooftops. Since then, Charlot said, the neighborhood’s social fabric has been frayed, along with its infrastructure. The St. Claude strip is largely dead now; most of the commercial buildings are now vacant. There is no good grocery store nearby, she said. No pharmacy. The streets are in disrepair, and traffic lights are frequently out.
After group held charrettes with residents, asking how to bring back a community flooded into ruin. Like people from other heavily flooded neighborhoods, Lower 9 residents are wary of long meetings that result in nothing.
Five years ago, the Lower 9th Ward Homeowner’s Association published an analysis of the 17 plans that had been created in the neighborhood, followed by limited or no implementation.
If, this time, the follow-through and resources were real, you’d get the attention of Charlot and other Lower 9 neighbors interviewed. Beautification efforts, for instance, might make a dent in crime, Charlot suggested. So might other investments, in people, through increased family resources and mental health services. In short – she’s open to ideas.
Though Charlot was not familiar with the proposed RTM project, she wondered why a big-dollar software package was necessary to identify conditions that neighbors already know. Instead, city officials should repair already-identified issues, she suggested. “We need streetlights, we do,” she said. She , noting that she is also open to more police presence in the neighborhood, though she’s aware of the dangers of over-enforcement.
That concern is shared by Omojola. She believes that, in high-poverty neighborhoods, the program could prompt targeted crackdowns.
“The first thing that jumped out at me about Risk Terrain Modeling is that it could really lead to over-policing in areas that are already under-resourced,” Omojola said. “In New Orleans, that means Black communities, and other communities of color.”
Crime analyst Jeff Asher, though a proponent of tools like RTM, also echoed skepticisms about the city’s ability to utilize the technology consistently over time. “Is the city able to do the stuff beyond the enforcement action, so it’s not just a short-term effect, but that there’s long-term benefits for everyone?” asked Asher, who has worked with the New Orleans City Council and former Mayor Mitch Landrieu.
Using data to fight New Orleans crime – with varying results
It’s not the first time that New Orleans has tried to drill down on data to improve public safety. Former NOPD Superintendent Richard Pennington, along with consultants John Linder and Jack Maple, launched “hotspot policing” here in the mid-1990s. Under Pennington, the city’s murder rate was cut in half, to the point where the chief sometimes got standing ovations when he entered local restaurants.
But their work also raised concerns about civil-rights violations.
Thirty years ago, as they began that work, the trio crunched stats on crime, discussed them in regular Compstat meetings with district commanders, and created “proactive patrols” – sometimes referred in neighborhoods as The Jump-Out boys – who would drive through blocks where most crimes were occurring and try to catch crime in action, sometimes by indiscriminately stopping and frisking and detaining “suspicious” people. The strategy went hand-in-hand with broken-window-era crackdowns on low-level violations.
More recently, in 2012, Mayor Mitch Landrieu’s NOLA for Life wide-ranging murder-reduction program began quietly partnering with Peter Thiel’s tech giant Palantir to use predictive policing software. They amassed data for six years on New Orleanians, to identify individuals with criminal histories who were at the highest risk for gun violence, as aggressors or victims, who were then targeted for group “call-ins,” to be both offered city services and warned about enhanced prosecution.
Landrieu – like Mayor LaToya Cantrell – expanded the city’s surveillance-camera network, which also has the potential to be misused.
Jeff Asher, who utilized the Palantir software for Landrieu’s administration between 2013-2015, cautioned that – like every tactic tried in New Orleans in the past — RTM alone is not a magic bullet. “It’s not, like, the game changer,” Asher said. “It’s not the thing that ends crime in New Orleans. It’s a tool amongst other tools.”
Plus, even programs that have been successful elsewhere can fail when they hit the ground here, Asher said. “New Orleans is just, sort of, New Orleans,” he said. “So things that work in other cities sometimes don’t work here.”
***
Democratizing data
By design, the entire RTM process must be guided by communities, NOPD and city officials, nonprofit organizations, and businesses. “What I want to do is get this data democratized, pushed down to the street level, and then use that data to drive decision-making,” Gaudin said.
Still, it remains to be seen how much input Brandi Charlot or other neighbors will have into RTM’s investments and interventions.
In Newark, the professors who came up with RTM established the Newark Public Safety Collaborative, which engaged with community organizations – primarily nonprofits and community groups – to use RTM data to direct their work, said Alejandro Giminez-Santana, a professor at Rutgers and the co-executive director of the collaborative. They worked with code enforcement to temporarily shut down corner stores where crime was concentrated. They also helped a local utility install floodlights in front of corner stores, for a nominal, $20, charge to store owners.
The software also guides the location of a beautification program, public-awareness campaigns to reduce auto thefts and robberies near ATMS, and a “safe-passage” program, where members of the Newark Community Street Team are deployed around schools and bus stops to intervene in potential conflicts.
Gaudin’s vision for New Orleans differs. He wants the whole of the city government – not just a handful of non-profits – to be responsive to the solutions developed by the RTM process, and to reduce the burden on NOPD.
“Taxpayers are paying for the Department of Public Works, the Public Health Department, the Coroner’s Office, the public school system, and code enforcement, as well as the City Council itself, which holds the purse strings of many of these things that we haven’t been backing the NOPD with,” Gaudin said.
Omojola, with Vera, supports targeted, non-law enforcement investments in areas where police make frequent arrests. But she would prefer to see the Risk Terrain Modeling pilot under the auspices of the city’s health department, which is working to implement its own violence-reduction strategy.
Gaudin, a true believer, says that RTM’s strengths make up for the blindspots he carries from his profession. “I’m limited to some of the dumb ideas that a prosecutor would come up with,” he said. “This concept, though, is not. This concept is wide open to creativity and things that are well beyond the normal law-enforcement scratchpad.”
Yet some wonder whether Gaudin’s own grand ambitions for RTM – to push boundaries and to find solutions beyond law enforcement – may be limited because RTM is being guided by his office, where trained prosecutors spend their days focused on prosecuting crime.
“When you’re a hammer, everything is a nail,” Omojola said.
Katy Reckdahl contributed to this story.
This article originally published in the November 20, 2023 print edition of The Louisiana Weekly newspaper.
