Chad Topaz

Applied Mathematician and Data Scientist Chad Topaz

During the early days of the covid-19 pandemic, New York City Mayor Bill de Blasio worried about the spread of the virus in city jails. To reduce that spread, the mayor announced that people who were not likely to commit new crimes would be released from Rikers Island, NYC’s main jail complex.

When he heard about Mayor de Blasio’s decision, Applied Mathematician Chad Topaz, a data scientist who had read about violence and other problems at Rikers, decided to use his math and statistical skills to study the success or failure of this effort.

Would the people released from Rikers in March, 2020, be more or less likely to end up back in jail by the end of the year, when compared with people released from the jail in March of other years? Topaz and his colleagues at the Institute for the Quantitative Study of Inclusion, Diversity and Equity (QSIDE, a nonprofit the mathematician co-founded with his husband, Jude Higdon) decided to figure this out.

As a first step, they gathered publicly available data on people released from Rikers. This turned out to be a difficult process: the information they found was not always complete or accurate. “We would uncover class after class of discrepancies.”

Topaz and his colleagues cleaned up the data, creating a more accurate database than the one available online. Then the researchers analyzed the data they had gathered, using different statistical tests and models.

The team also spoke with various experts, knowing that they needed to better understand the information they had found about each inmate. “These are people: they’re not just numbers in a spreadsheet.

What did they find? Descriptions of inmates’ race were often left out of the publicly available data. And in a city with a very diverse population, the data on race for Rikers inmates is organized into just three categories: black, Asian, or “unknown.” (The U.S. Census uses eight categories to describe the population of New York City by race.) “The way we keep and talk about data has the potential to directly harm the people described by that data,” says Topaz, who continues to wonder why the Rikers list categorizes people this way.

What Topaz describes as “the main headline of the project” investigating the success of the 2020 release of prisoners during the COVID-19 pandemic is this: “Why are [NYC officials] not letting people out of Rikers all the time?” The people in charge of the March, 2020 release were able to figure out which people locked up at Rikers were not likely to commit new crimes. The people they let out of jail were less likely to end up back in jail than groups released from Rikers at other times. Why were these people in jail in the first place? And why aren’t these kinds of releases happening regularly?

Perhaps part of the reason that Topaz co-founded QSIDE and focuses some of his mathematical research on social justice is that Chad “grew up as a gay kid in the ‘80s and ‘90s.” He came out during his sophomore year of college, deep into the AIDS epidemic. Sometimes, “society tells you how bad you are … that you’re not a worthwhile person.”

Even now, “it’s not awesome” to be gay in the field of mathematics. Math “seriously has a lot of issues with equity, diversity, and inclusion. It can be a very sexist, homophobic place.” But “there are many people working on this” and Topaz and his colleagues are encouraging their professional associations to make more equitable, welcoming decisions.

“The thing that I love about applied mathematics is that I never have to choose: I can indulge my multiple interests.” In graduate school, Topaz studied pattern formation in fluids. In later years, he looked at the ways groups of creatures move. Anchovies, for instance, “make a big tornado,” swimming together in very large numbers. “If you’re a fish, you have some behavioral rules that your brain wants you to do: go near other fish but don’t get too close to them.” Moving in large schools makes it easier for fish to find food, reproduce, and scare away predators.

Flocks of birds like starlings and swarms of insects like locusts move in large groups for similar reasons. How do these different animals organize their movement? “When you study these groups, you have a ton of data.” For instance, if you look at flocks of starlings, you might note the position of each bird in three-dimensional space and also its velocity (its speed and the direction it’s moving in). That gives you six data points per bird. If you have videos of the flock’s movement, it provides you with tens of billions of pieces of data. How can people studying these groups make sense of that information?

“I was talking about this problem with a friend of mine” at Macalester College, where Topaz taught in the past. Mathematician Lori Ziegelmeier, a topologist, “said ‘Chad, why don’t you think of your data topologically?’” “[She] changed my life by teaching me this stuff.” Together, they imagined each starling as a point in six-dimensional space, connected to other nearby birds/points. What kinds of objects did the groups of points form? How many pieces and holes would these objects have? This way of thinking about large groups of animals is a “powerful way of analyzing big and complicated data” describing their movement.

Topaz has recently worked on honeybees “with amazing people at the University of Colorado.” The bees “go out, get food, and share it by spitting it into each others’ mouths,” a behavior called trophallaxis. Topaz is using topology to better understand the clusters the bees form and the ways they spread out as they feed each other.

The mathematician also spends a lot of his time working with his students at Williams College. Not satisfied with change he can make as a professor at one private college, however, he and QSIDE have looked into what prevents members of minority groups from attaining degrees in STEM at the rates white males do. They found that it’s not for lack of interest, skills, or prior academic success. Instead, they found that withdrawing from or failing just one STEM class towards the beginning of college significantly reduced students’ chances of finishing school with a STEM degree “in a way that differentially impacts students of color.”

In addition, Topaz and his colleagues at QSIDE have analyzed how well the work of female artists and people of color is represented in major museums’ art collections. Their conclusion? “The time is long past for major art museums to become activist collectors, emphasizing the work of women and artists of color … to address historical underrepresentation and bias.”

Based on the results, data journalist and artist Mona Chalabi created images to represent the demographics of the data set, scaled down from over 10,000 artists to 100 for the purpose of visualization. In this case, 88 of the artists would be men (75 White, eight Asian, three Latinx, one Black, and one man of another race/ethnicity). Figure courtesy of Mona Chalabi.

And Topaz and QSIDE reached out to the people in charge of the Rikers Island data, prosecutors, and defense attorneys, trying to understand why there were so many errors and omissions in the public information on inmates and why the data on race is organized into just three categories. These efforts did not lead anywhere, so Topaz and Higdon wrote an Op-Ed for the New York Daily News. Never one to consider the people he studies as abstract data points, Topaz uses his mathematical skills to promote positive change.