Case Study 06

The Data Analysis That Exposed Racial Profiling in a Neighborhood Watch App

Publication: The Georgetowner · Role: Investigative reporter · When: 2014–2016
Investigative journalism Data analysis Original reporting National media pickup
3+
Months of investigation before publication
1,000+
GroupMe messages analyzed to identify the pattern
4
National media outlets that picked up the story

The story

The Georgetown Business Improvement District operated a neighborhood watch program built around a private GroupMe group chat. Residents and business owners used it to flag suspicious activity in the area — describing people, sharing locations, alerting others to follow up. It was framed as a community safety tool.

I obtained access to more than 1,000 messages from the group and analyzed them. The pattern that emerged was stark: 72 percent of the people described as suspicious were Black. Only 16 white people were flagged across the entire dataset. The app wasn't functioning as a neutral safety tool. It was systematically targeting Black people in a neighborhood where they were not the majority of residents.

72%
Of people flagged as suspicious were Black
Across 1,000+ messages analyzed
16
White people flagged in the entire dataset
By comparison

The BID disputed the findings. Critics I spoke with did not. The story ran in The Georgetowner in October 2015 and set off a chain of events: national media pickup, public pressure on the BID, and — within weeks — the BID shutting the app down entirely.

The reporting

The story required obtaining the messages, cleaning and categorizing the data, identifying the pattern, and then reporting it out with sources on both sides. The BID was given the opportunity to respond; their response was included. Residents, civil liberties advocates, and community members provided context for what the numbers meant.

The data didn't require interpretation to be damning. The job was to report it accurately, present it clearly, and let it speak.

The Georgetowner was a small neighborhood publication. This was not the kind of investigation that typically originates at a hyperlocal outlet, which made the pickup pattern that followed notable: the Washington Post ran a version of the story without attribution, CBS This Morning covered it, Fox 5 reported on it, and NPR picked it up. I was asked to appear on CNN but was bumped before the segment aired.

The BID responded to the coverage by kicking The Georgetowner off the GroupMe app. The app itself was shut down shortly afterward.

How it unfolded

Aug 5, 2015
Original story publishes

The Georgetowner runs the initial investigation online and in print: data analysis of 1,000+ GroupMe messages, findings on racial disparity in flagging, BID response, and community reaction.

Aug 2015
Fox 5 coverage

Fox 5 covers the story in the weeks following publication. The BID responds by removing The Georgetowner from the GroupMe app.

Aug–Sep 2015
BID retaliation

Public pressure builds. The Georgetown BID removes The Georgetowner from the app in response to the reporting.

Oct 2015
Washington Post publishes

The Washington Post runs their own version of the story. CBS This Morning and NPR coverage follows shortly thereafter. CNN books me for an appearance; the segment is bumped before air due to the Israel-Palestine conflict.

Nov 2015
BID shuts the app down

Under sustained public pressure, the Georgetown BID discontinues the GroupMe program entirely.

The articles

The Georgetowner · Original investigation
Critics claim racial bias in Georgetown's digital crime prevention program
The original investigation. Data analysis of 1,000+ GroupMe messages revealing that 72% of people flagged for suspicious behavior were Black, with responses from the BID and critics of the program.
Read the article
The Georgetowner · Follow-up
GroupMe media frenzy
Follow-up coverage of the national media pickup: Washington Post, CBS This Morning, Fox 5, NPR, and the broader public reaction to the original investigation.
Read the article
The Georgetowner · Follow-up
BID shuts down public safety app GroupMe
The final piece in the series, covering the Georgetown BID's decision to discontinue the program entirely in the wake of the investigation and subsequent public pressure.
Read the article
National pickup
Washington Post, CBS This Morning, Fox 5, NPR
The story was picked up by four national outlets. The Washington Post ran a version without attribution. CBS This Morning, Fox 5, and NPR all covered the findings independently.

What this work demonstrates

The story worked because the data was airtight. No anonymous sources, no agenda — just a dataset, the work to understand what it showed, and reporting precise enough that the findings held up under national scrutiny from four outlets.

It also demonstrates something about consequence. The BID shut the app down. That's a direct, traceable outcome from a piece of journalism that originated at a small neighborhood publication, and the fact that it crossed into national coverage without losing the rigor of the original analysis is something I'm proud of.

The skills that made this story work — finding the pattern in a dataset, structuring a complex finding clearly, writing with enough precision that nothing can be picked apart — show up in every piece of technical content I've written since. The subject matter changed. The standard didn't.