STORY: Olin Students Fighting Mass Incarceration through Data Analysis
September 1, 2023
ADE project building tool to examine racial disparity in police stops
Within the Affordable Design and Entrepreneurship (ADE) program, a team of Olin students have been building data tools to help individuals avoid possible conviction and incarceration due to unlawful traffic stops resulting from racial profiling.
Erhardt Graeff, associate professor of social and computer science, has been collaborating with the Committee for Public Counsel Services (CPCS), Citizens for Juvenile Justice, and other advocates on this project since 2020. Using data describing every citation issued to a driver in Massachusetts in recent years, Graeff and his student research team are writing Python software to calculate racial disparities in who is stopped and cited given the demographics of the surrounding area.
Photo of summer research team: Anmol Rattan Singh Sandhu ’25, Erhardt Graeff, and Bethany Costello, a recent alum of Wellesley College who cross-registered for Erhardt’s ADE course last year.
“Pulling someone over is a discretionary police activity and a rubric that should be applied equally across all races,” says Graeff. “However, the data shows that’s not the case, with folks who are or appear to be Black more likely to get pulled over, which can lead to incarceration because of additional charges, such as drugs or firearms. One of the ways to address mass incarceration is to use data to more effectively keep folks out of prison.”
In addition to traffic stops, people of color are incarcerated at higher rates in America; a recent study showed that Black Americans are incarcerated in state prisons at nearly 5 times the rate of white Americans. In addition, the disparity between Latinx and white people who are incarcerated is highest in Massachusetts, with a differential of 4.1:1 (compared to 1.3:1 nationwide).
Leveraging Long Motions for Equality
In 2020, a Massachusetts Supreme Judicial Court decision in Commonwealth v. Long made it so that public defenders could more easily pursue motions to dismiss evidence under the Constitution’s Equal Protection clause. These “Long motions” lay out the potential to use data from police records to indicate that race may have been a factor in the decision to pull someone over.
This decision sparked an idea by Joshua Dankoff, director of strategic initiatives at Citizens for Juvenile Justice, about creating a data tool that could be used by public defenders and their clients to argue Long motions to exclude evidence obtained during unlawful stops. Dankoff pitched the idea to Graeff, which kicked off five semesters and two summers of research and design work by Babson, Wellesley, and Olin students on the team. After this initial groundwork, significant development has been made this summer by Graeff’s two 2023 research assistants: Anmol Rattan Singh Sandhu ’25 and Bethany Costello, a recent alum of Wellesley College who cross-registered for Erhardt’s ADE course last year.
Sandhu and Costello have been writing and documenting new code to develop a reliable data pipeline that ensures the quality of the compiled data and resulting statistics. By the end of the summer, they hope their code can automatically generate statistical reports that CPCS’s Strategic Litigation Unit can easily add to their new police accountability database, which attorneys can query for information relevant to their cases.
“I’m a computing major at Olin, so data analysis and working with Python is really interesting to me,” says Sandhu. “We were able to do a lot of independent coding to figure out how we can solve new problems that come about throughout the project. It’s super hands-on because we have a lot of control over what’s happening and the impact the data tools can make in the real world.”
“This CPCS project was a great way to combine my skills in data science with my passion for social justice,” says Costello, who was a political science major with a computer science minor at Wellesley. “I want to be able to pursue a career combining those two things and working on this project for the past two years and learning data tools like NumPy and Pandas has been a great experience.”
“The students have built the whole system, including working with defense attorneys and other legal advocates to understand their needs,” says Graeff. “As with any ADE project, this is community-designed work, and the students want to make sure they engage with the end user and really understand the problem. Not only are they learning more about design cycles and the legal system, but they’re also stretching their understanding of the roles and responsibilities of what it means to be an engineer or technologist in a problem-solving context.”