Teaching data journalism

The opioid reporting project was conceived by Tracy Breton, a Pulitzer Prize-winning journalist and a professor of the practice who has taught journalism courses at Brown for more than two decades. In recent years, Breton has become increasingly interested in data journalism — the use of large datasets in developing news stories or bolstering facts. That interest is driven partly, she says, by the charges of “fake news” increasingly leveled at the American news media. 

“I really felt like it was more important than ever for us as students and journalists to produce stories that were powerful, fact-based and data-driven, that no one could contest as being fake,” Breton said. 

The concept of using data to unearth new truths on pressing issues meshed ideally with the pervasive drive of Brown students to solve problems in society. The commitment of Brown student and faculty scholars to joining forces across fields of study, a signature of the University’s Open Curriculum, didn’t hurt either.

In 2018, Breton and a group of independent study students produced a series of stories on elder abuse in Rhode Island. The team pored over every criminal case in the state since 2000 involving elder abuse and found that few people ever serve jail time for abusing and exploiting elders. Their findings, along with personal stories of those who had suffered abuse, ran over nine days in the Providence Journal and was nominated by the newspaper for a Pulitzer Prize.

For several years, Breton had been thinking about bringing that same data-driven approach to reporting on the opioid epidemic, but she knew the datasets involved would be daunting. Sifting through them would likely require expertise beyond what she or her journalism students could muster. So in early 2020, Breton reached out to Ugur Cetintemel, chair of the Department of Computer Science. 

Breton pitched Cetintemel on the idea of a course, cross-listed with the English and computer science departments, that would bring data scientists and journalism students together to tackle the opioid issue. The course would take a year or two to get on the books, but Breton believed there would be plenty of interest. 

“He looked at me and he said, ‘Why do we have to wait another year or two to do this?’” Breton said. “‘I’m teaching a class this semester called CS for Social Change.’” 

A key component of that class, which Centinemel has offered for the last five years, involves student teams partnering with local organizations to help with technical challenges they face. Students have tackled a range of projects. One team worked with a criminal justice advocacy group called Open Doors on a data platform to study correctional data. Another worked with a nutritionist to develop an app that provides healthy food incentives. 

“I told Tracy that I was teaching this class, and that we can have one team of students working with her on this series of stories,” Cetintemel said. “She liked the idea, and that’s how this got started.”

Data and stories

Triedman, a computer science and history concentrator, was one of the five CS for Social Change students who joined the project, working with 12 journalism students in Breton’s Reporting Crime and Justice class. The work involved easily spilled the banks of the semester-long classes. The team ended up working together for nearly two years. Triedman was one of several students who continued working on the project even after he graduated.

The datasets used in the project were often enormous and complex, he said. Data from the DEA, for example, detailed every step of every legal opioid transaction in the U.S. for over a decade — from importation of the opium poppy, to its processing in a pharmaceutical plant, to its sale in a local pharmacy. The dataset was hundreds of gigabytes in size with hundreds-of-millions of rows of information, Triedman said. 

“You can't load that onto any normal computer, let alone open it up in Excel and do the normal filtering you would want to do,” he said. “It takes some specialized skill to work on those datasets and to know which tools to use.”

By sifting through that DEA data, Triedman was able to uncover the outsize role that Coventry’s Rhodes companies played in distributing opioid products nationwide, as well as which pharmacies in Rhode Island were handling and selling the most opioid products.

Other datasets provided other insights. Judiciary data enabled students to see evolving trends in how the criminal justice system has treated narcotics cases over the past 30 years. Data from the Rhode Island Board of Medical Licensure and Discipline helped Li Goldstein, currently a Brown senior, to tell the story of Dr. Christopher Huntington, a Providence orthopedic surgeon who wrote illegal prescriptions for thousands of opioid pills before committing suicide in 2013. 

Breton said she was careful to make sure the students were making good use of data, but not neglecting the human side of the story.

“I wanted [the stories] to be data-driven, but I also know that readers can have their eyes glaze over if all we're looking at is numbers and charts and graphs,” she said. So she instructed students to include “faces in the story that help explain what the numbers are showing.”

For a story that mapped out three years of overdose deaths in Providence, student reporters Olivia George, Marina Hunt and Colleen Cronin crisscrossed the city to talk to people whose lives had been touched by the epidemic. 

“Mapping the locations of fatal overdoses in the city was both helpful and overwhelming in that it provided us with a striking visualization of the persisting scale of this public health crisis,” said George, who is currently a senior. “But in this work we can never lose sight of the fact that each number, each statistic represents a human being, a community member. I did not want us to simply reduce someone to a pin on a map.”

George said that speaking to individuals who have been impacted by the epidemic was a deeply moving experience for her personally.

“What a gift it is to be present with people. I hope that our reporting, even in some small way, speaks to the sense of gratitude that we have for all those who shared their time, energy, experiences and expertise with us," she said.

A formative experience

Triedman considers his opioid reporting work to be the capstone experience of his time at Brown. 

“I think that this project took some of the researching and digging instincts that I had developed over my time as a history concentrator and combined them in a really interesting synthesis with the quantitative and user design-skills that I had acquired over the course of my time as a CS concentrator,” he said. “It provided this amazing fusion of those two disciplines that really rewarded both areas of expertise.”

Triedman is currently working as a programmer for the Wikimedia Foundation and says he hopes to continue doing journalistic work in the future.

The project had a major impact on George as well. Last summer, she was a reporting intern at the Virginian-Pilot, a daily newspaper in Norfolk, Virginia. She says her work with Breton made that internship possible and helped provide direction for her career moving forward. 

“Working with Professor Breton, especially on this project, has really solidified in my mind the importance of local reporting,” George said. “I have become rather married to the idea, not only of being a journalist, but really wanting to cover a community in which I live.”

Breton sees projects like this one as a model for local investigative reporting. 

“With all of the downsizing and layoffs at newspapers around the country, I do think that the future of local investigative reporting lies largely in the hands of seasoned reporters who are willing to teach and work with advanced journalism students at our colleges and universities,” Breton said.

She said she plans to continue combining data science expertise with journalism to tell stories that are important to Providence and the Rhode Island community.  

“The truth is very hard to find and matters now more than ever. We need to continue to dig deep — with on the ground interviews, public records research and data science — to tell the untold stories that need to be told.”