The "Avoiding the Ghetto" app

By John Roman :: January 31st, 2012

Share

 

Recently, Microsoft received a patent for a new app that allows users to avoid dangerous areas as they walk around town. The patent application states, “As a pedestrian travels, various difficulties can be encountered, such as traveling through an unsafe neighborhood or being in an open area that is subject to harsh temperatures.”

In an interview with WAMU here in Washington, DC, University of Maryland criminologist Charles Wellford does an excellent job summing up the case for and against the app. Wellford says, “If the app does provide accurate information, it could be useful in helping people decide whether they want to adjust their behavior, such as walking down a different road.”

Seems innocuous enough. Yet, the app has launched a thousand reactionary headlines. Critics are calling it “appalling” and warning that it could lead to the kind of red-lining that mortgage lenders used in the 1950s and 1960s to isolate African Americans in less attractive neighborhoods. Pretty important stuff for one lonely app.

So, should we be concerned about the “avoid the ghetto” app?

I vote no. And yes. No, because no app can really do what the Microsoft app claims to do, as I show below. And yes, because in opposing this app, well-intentioned people perpetuate the (incorrect!) stereotypes they seek to defend against.

Let’s consider how a ghetto app would work. The idea here is to keep you safe as you walk about town.  My work takes me to some of the country’s most dangerous cities, and I have often thought it would be nice to know if a particular neighborhood was safe for me to explore. If I lived there, I might like to know where it would be okay for my kids to go without me.

In many people’s minds, an unsafe area looks disorderly—groups of potentially dangerous folks loitering about and lots of abandoned cars, graffiti, litter, and so on. That image might hold for cities in the northeast, but if you were to walk through a high-crime public housing complex in Portland, Oregon, you would not see any of those physical cues.

So, what actually determines whether a place is too dangerous to traverse? How much crime occurs there, of course. Let’s walk through Washington, DC, and see where the ghetto app should tell us not to go. More than half of all blocks in DC report fewer than 5 crimes in any given year, meaning there is a crime every couple of months. So, there’s little crime to avoid in most places.

Average Crime by Block in Washington, D.C., 2000-2009

 

Source: DCPI analysis of Part I crime incident data from the Metropolitan Police Department, Washington, DC, January 2012

But some places in DC experience quite a lot of crime, and these are fodder for the app. Logically, however, the type of crime matters a lot in determining whether an area is safe to walk through.

Most crime is property crime. Looking at the most serious crimes (what used to be called Part 1 crimes) reported to the FBI in 2010, theft, burglary, and motor vehicle theft accounted for more than three times as much crime as serious person crimes (murder, rape, robbery, and assault). There were 36,000 Part 1 crimes in DC last year, but only about 8,000 were person crimes.

Now, the app is designed to guide our walk (so forget motor vehicle theft) through an area where we do not live (so forget burglary) or own a business (so forget theft), and where we probably don’t know anyone (if we did, we probably wouldn’t need the app!). That means what we care about avoiding is being hurt or robbed by a stranger.

First and foremost, we want to avoid being killed. While DC’s murder rate is declining rapidly, it’s still the 4th highest among big cities in America.   We would want to focus on recent homicides, of which there were 108 in 2011. Once we exclude the murders where the killer and victim are somehow connected (parents killing infants, husbands killing wives, guys in a crew killing each other, friends or coworkers killing each other, etc.), very few murders are left for the app to map.

We would like to avoid being raped too. But, there are fewer than 200 rapes a year in Washington, DC, and in most cases the victim and offender know each other. So, again, there’s not much data to go into the app.

That leaves assault and robbery (taking something from a person by force or threat of force).  There are about 4,000 assaults and 3,000 robberies a year in DC, so these seem like good candidates for our app.

Checking out a map of 2009 data, the part of DC with the most assaults is in the 3rd police district, an area known as Adams Morgan and Columbia Heights. (Maps of previous years’ data show the same hot spot.) These places are not the poorest in the city, nor are they the areas with the most minorities. What makes Adams Morgan and Columbia Heights so dangerous? That’s where the bars are heavily clustered.

Where are robberies most concentrated? Same place! And, within that place, the “hottest” hot spots are near Metro stations and along the busiest commercial corridors (where the most bars are).

Maps of Robbery Hot Spots by Police District in Washington, D.C., 2000-2009


Source: DCPI analysis of Part I crime incident data from the Metropolitan Police Department, Washington, DC, January 2012

In other words, walking through a poor or minority neighborhood doesn’t automatically make you more susceptible to serious personal crime. Walking late at night through a heavily populated commercial area, especially one with lots of impaired people, does.

So, a word to those who demagogued this app: your assumption that the kind of crime a person walking through a city would most want to avoid would naturally cluster in poor, minority areas is wrong.  Please let go of this stereotype.

And to Microsoft I say this: I really don’t think I’ll pay 99 cents for an app that tells me I am safer if I avoid bars and public transportation. But good luck anyway!

2Comments

  1. Rob  ::  1:59 pm on February 1st, 2012:

    Also interesting to note – Georgetown U, American U, Gallaudet U, and Catholic U all appear to be shaded in the purple (more than 25 crimes per year). This probably makes sense, but without that context, these look like high crime outliers in otherwise low crime areas.

  2. Crowd Sourcing Police Work: Bad Idea  ::  10:22 am on July 16th, 2012:

    [...] citizens to avoid places where their likelihood of criminal victimization increases (spoiler alert: it won’t work). That, however, is far less dangerous than ill-advised apps allowing private citizens, at the [...]