Hi, I’m redshiftzero. I’m interested in security, data analysis, and machine learning.
Professional Experience
Freedom of the Press Foundation
SecureDrop Lead Developer, July 2017 - present
Ford-Mozilla Open Web Fellow, September 2016 - June 2017
Center for Data Science and Public Policy
Postdoctoral Researcher, September 2015 - September 2016
Data Science for Social Good Technical Mentor, June 2016 - August 2016
Data Science for Social Good Fellow, June 2015 - August 2015
Teaching
University of Chicago, PPHA 30530: Computation for Public Policy, Winter 2016
Education
University of Chicago, Ph.D. Astrophysics, completed June 2015
Projects
SecureDrop
SecureDrop is an anonymous whistleblowing platform for journalistic sources to use to send files and messages to news organizations. It is currently widely deployed in dozens of news organizations, including The New York Times, The Washington Post, the Associated Press, the USA Today, and more. You can see the full list of deployments in the SecureDrop Directory. See how it works in this short YouTube video.
Code: GitHub repository
Role: Maintainer (2016-present), lead developer (2017-present)
OpenOversight
OpenOversight is an interactive web tool and accountability platform that makes it easier for the public to identify police officers, including for the purpose of complaints. It relies on crowdsourced and public data to build a database of police officers in a city. Using OpenOversight, members of the public can search for the names and badge numbers of police with whom they have negative interactions using the officer’s estimated age, race and gender. Using this information, the OpenOversight web application returns a digital gallery of potential matches and, when possible, includes pictures of officers in uniform to assist in identification. It is currently deployed on the Chicago, Oakland, Berkeley, and University of California Berkeley Police Departments.
Code: GitHub repository
Media: Citylab, Chicagoist, Washington Times
Role: Began project (2016), primary developer (2016-present)
Predicting Police Misconduct
This project is a data-driven Early Intervention System (EIS) for police departments. The system predicts which officers are at the highest risk of an adverse interaction with the public using machine learning models trained on officer-level historical data from the department. It is currently deployed in Charlotte-Mecklenberg Police Department.
Paper: “Identifying Police Officers at Risk of Adverse Events” KDD 2016 (PDF)
Code: GitHub repository
Media: Mother Jones, FiveThirtyEight, NPR, wsoctv
Role: Primary data scientist while a postdoc (Sept 2015- Sept 2016), technical mentor to two teams of Data Science for Social Good fellows working on the models (Summer 2016)
Predicting Building Code Issues
As part of the Data Science for Social Good Fellowship in 2015, my team and I built a machine learning model that predicted which properties in the city of Cincinnati were likely to have building code violations, used for allocating limited city resources.
Code: GitHub repository
Media: Wired/Backchannel
Role: Member of data science team