The Parlour in downtown Durham is a happy place where the line for ice cream often flows out the door. Children enjoy their cones on the plaza outside and chase each other around Major, the city’s landmark bull statue, as their parents look on.
Those parents probably aren’t thinking about a claim that they’re 16 times more likely to be robbed and eight times more likely to be killed on that plaza than the national average. At least, that’s what it says in Crime and Place, a new iPhone app that says it’ll tell you the risk of crime for your location.
Crime and Place calculates the likelihood of murder, rape and other crimes by distilling data from the FBI’s Uniform Crime Reports into a score from 1 to 10 for every census block group in the United States. The app uses maps and a “Crime Compass” that tells users not only how safe they are, but essentially how to escape high-crime areas.
That approach bothers some policy experts, including Robert Korstad, a Duke professor of public policy and history who has researched Durham extensively, and Martine Aurelien, policy fellow at the N.C. Justice Center. They say the app has useful features but fails to account for daily and seasonal fluctuations in crime. They’re also concerned that it can reinforce stereotypes and fails to account for users’ race and gender, which influences how likely they are to be the victim of crimes.
Sign Up and Save
Get six months of free digital access to The Herald Sun
Korstad called the app “both impressive and horrifying.”
For just $15 plus $5 for a year’s worth of data, Crime and Place will tell you where the crime is.
As you drive through a city, the app’s compass changes colors, giving you an escape route from the bad guys. Drive around Durham, for instance, and you’ll see it go from green (“low”) to yellow (“average”) to light orange (“elevated”) to red (“high”).
The app’s user manual offers some advice for road trips: “Try to stop for gas, meals, etc. when the compass shows green. Avoid stopping when the compass shows red.”
The Crime Compass shows high levels around Durham Tech. Most of Duke’s campus is in a yellow “average” zone, but walk to Ninth Street for brunch at Elmo’s and the crime level becomes “elevated.” That’s still a safer meal than in the heart of downtown at Pompieri Pizza or The Pit, where it’s “high.” Hope Valley Country Club is in a “low” region, but that isn’t true for all of Durham’s wealthy neighborhoods — crime is “elevated” at Croasdaile Country Club.
If you want the app to keep you safe without having to watch it constantly, you can set a threshold for notifications that inform you, “Entering high crime area!” whenever the rating rises above your chosen number.
Crime and Place’s creators say its meticulous data modeling makes it better than other crime apps. For each area, it calculates a separate number for murder, rape, robbery and assault to create an overall crime score. It also tallies a property crime score using burglary, larceny and auto theft. It shows those scores relative to the national average, which is 4.0. Each additional point on the scale doubles the likelihood of crime relative to the 4.0 national average.
So at the Durham Bulls’ ballpark and the Durham Performing Arts Center, the 8.1 robbery rating means you’re more than 16 times more likely to be robbed than the national average.
Crime and Place’s lead developer, Jolly Salehy, said the app is primarily intended to inform people visiting unfamiliar cities or looking to buy a home, but he acknowledged it could change how people view their own hometowns.
“A lot of us had similar stories of business trips where you’re sitting in a rental car or you’re in an unfamiliar city and you end up somewhere that is unsafe, and then you start to get that panicky feeling,” Salehy said. “Living in a place for a while, you get a feel for where is generally safe and where is generally unsafe, so it possibly has limited use in those cases, although we have had some people that end up being kind of surprised.”
The maps on the app show how arbitrary the differences between its labels can be. The border between “average” and “elevated” crime is roughly the 50-yard line at Wallace Wade Stadium. Visiting teams looking at the compass from their locker room would feel less safe in Durham than Duke’s players in the home locker room on the other side of the stadium.
The darkest shade of red in Durham is around Durham Technical Community College and the McDougald Terrace public housing community southeast of downtown. The crime score is as high as 7.3, and murder and robbery ratings are 8.3. The Crime Compass calls it “high.” That red region borders N.C. Central University, a historically black public school, and roughly correlates with the old Hayti neighborhood when the city was segregated. Today, the area is home to many in Durham’s low-income housing.
“Looking at that app and understanding the history of Durham, I think it just helps to corroborate an unfortunate narrative that has already been in place,” Aurelien said.
Other crime apps have been driven out of existence after being labeled as racist. SketchFactor, a crowdsourcing app in which the public could assign subjective ratings to points on a map and submit comments, was rebuked upon its launch in 2014 and is no longer available.
Aurelien said Crime and Place is more reputable because it uses pure data, but said it could provide more nuance to help overcome stereotypes. If people could see the crime rate in a community is decreasing, for example, they might be more encouraged to shop, eat or even buy a house there.
“Sometimes people aren’t just interested in knowing and understanding whether crime is prevalent but also understanding if it’s growing, decreasing or remaining stagnant,” she said.
Salehy said the app’s creators talked about racial stereotypes before launching Crime and Place in 2016 and called it a “touchy subject,” but decided there was no harm in repackaging public information.
“That was one of the things we kind of struggled with early on — how is this going to be interpreted by the public?” Salehy said. “At the end of the day, all we’re doing is facilitating visualization of some data. We can’t control how our customers choose to interpret that or the decisions that they make based on that.”
The FBI’s Uniform Crime Report (UCR) — the source of the app’s data — also has been criticized for being incomplete. The data is self-reported by localities, who may use different definitions and timetables for reporting incidents to the UCR.
Night and day
Aurelien noted that the app doesn’t change based on the time of day, but crimes do.
“There are some areas where there’s only a handful of petty crimes that occur during the day, but the mood completely changes at night,” Aurelien said.
The app’s ratings also do not shift by season even though some crimes are more common during the summer when more people are outside.
Korstad’s biggest qualm with the app is that it doesn’t evaluate how susceptible each user is to crimes. He said it would be useful if he could punch in demographics such as, “White college professor with a Ph.D. driving a big car that has plenty of speed and plenty of gas,” and get a customized prediction of his likelihood of being a crime victim.
Although the first sentence of Crime and Place’s methodology page states that it presents “well established, long term trends in criminal activity in the United States,” Korstad said there is a difference between criminal activity and crime rates. Criminal activity is not always caught or enforced by police, who often monitor and arrest people in minority communities at a disproportionately high rate.
“If I was interested in criminal activity around drugs, I know that one of the major places for criminal activity involving drug possession, drug sales and drug use is the Duke University campus,” Korstad said. “Because the police aren’t enforcing the law and they aren’t arresting or trying to arrest Duke students, it skews these numbers.”
That raises another issue that makes the app’s data incomplete: Crime and Place doesn’t include data on drug crimes at all. The seven crime categories it tracks make up Part I of the FBI’s Uniform Crime Reports, and drug offenses fall under Part II, making the app’s claim that it “paints a more accurate picture of overall crime levels in a given area” only partly true.
Hank Tucker covers public safety for The 9th Street Journal at Duke University.