Hoping to get approved for medical care? Land a job interview? Obtain a home loan? Qualify for a governmental program?
Chances are your fate will be determined by an algorithm — a mathematical formulation that you will never get to see and that you cannot control.
In this era of big data, algorithms have become an essential part of most institutional processes. But often even those commissioning them don’t understand the driving data points. And too often, the underlying data can be correlated with information on income, race and gender — discriminating against and harming communities of color and others.
“Our role as journalists is to hold power accountable. Black boxes cannot today escape a reporter’s scrutiny,” says Sandeep Junnarkar, director of the Data Journalism Program at the Craig Newmark Graduate School of Journalism. “It’s no longer sufficient to just interview officials and the people affected by an issue. We need to question how decisions were made, whether machines were used, and how those machines were trained.”
Junnarkar and his colleagues are working to empower the next generation of journalists to investigate these hidden forces, training them to use tools that haven’t yet made their way into most newsrooms.
The school’s Interrogating the Algorithm course, taught by BuzzFeed investigative reporter John Templon, helps students learn how to identify when algorithms are used, create public records requests to access their computer code, and create further requests to access the formulations’ underlying datasets.
The module also teaches third-semester data journalism students — many of whom will soon be heading into newsrooms where these approaches are novel still — how to speak clearly and simply about this work with editors and peers.
Every student who passes through the Newmark J-School program completes a core course in data journalism. For those interested in the investigative aspects of data journalism, additional courses cover advanced skills such as coding with Python and using machine learning and artificial intelligence to power investigations.
These are essential tools in an ever-changing field, says Junnarkar. He keeps his skills sharp through his work as a data journalism trainer at Bloomberg. There he teaches journalists to automate the acquisition of data from websites and employ optical character recognition to fully access scanned documents.
“At this point, I think it would be irresponsible of any journalist to write an in-depth story without at least considering the available data to see if it backs up their conclusions,” Junnarkar says.
Ultimately, data offers journalists the tools to uncover and document stories with high-impact on communities that might otherwise be marginalized, says Lam Thuy Vo, a data-journalist-in-residence at the school. It’s one of the reasons the school received a grant in 2019 from the Knight Foundation to have Junnarkar lead a program to train community journalists to report on and convey the ramifications of AI in an understandable and culturally nuanced way.
Vo often uses data to examine power imbalances where inequality and race intersect, for example by showing how gentrification ties to increased quality-of-life complaints and over-policing of people of color. Recently, her students published data-fueled articles examining housing insecurity and gentrification in New York City during the pandemic.
Vo says that her experiences as a child of Vietnamese refugees instilled in her a great awareness of how systems and the people at the heart of them can become oppressive. Data, she says, can illuminate how this occurs and empower those who often are unheard and deeply affected.
Data journalism skills are vital for journalists of color who can “bulletproof” their reporting of racism and inequality in stories that may face initial skepticism, she says.
“Data, especially in the past decade or two, has become one of the most powerful tools to wedge your editorial vision into newsrooms that might not immediately see it,” she says. “I want to give my students an instrument to say, ‘You can’t ignore me.’”
The opportunity to mentor the future journalists who most need this added firepower is a large part of why she chooses to teach at Newmark, says Vo.
“One of the reasons I’m here is because of the diversity of the student body, and because this school allows most of them to graduate with minimal debt,” she says. “I want to give them an instrument to come into the newsroom and convey, ‘I am irrevocably good. You cannot ignore it.’ And data is that instrument.”
Vo encourages students to pair rigorous data analysis with compelling human stories — what she calls “the full swing.”
“I try to ensure that at the heart of it all is still a good story,” she says. “Our job as data reporters is not just crunching the data but ensuring that we fully understand it and synthesize the findings in ways that are clear, void of jargon and informed by reporting.”
Junnarkar says that data journalists must help their audience understand how the news directly affects their lives.
“What I try to impart to my students is that it’s vital to humanize the data — to create ways for people to see how they fit into the news. Interactives that tap the data should provide insight into the audience’s own situations or experiences,” he says.
Students don’t need to be math whizzes to master now-essential data skills, the instructors emphasize. Junnarkar took challenging math courses in high school but never thought of himself as “math-inclined.” Vo was always more interested in writing and the humanities, focusing her studies on German and Italian culture and language.
Some of the program’s data journalism students have used optical character recognition and Python to uncover the NYPD’s lax responses to official misconduct . Others have built interactives highlighting the increased likelihood of exposure to gun violence among Black New Yorkers and the true financial impact of proposed congestion-pricing laws.
As advancements occur in data gathering, processing, and application, they play a burgeoning role in shaping peoples’ lives — and journalists must know how to recognize this and challenge its power and those who might abuse it, Junnarkar says.
“Data is the new gold. It’s the new oil,” he says. “All this data can be used, or it can be misused. So as journalists, we need to be aware of how to get a hold of it, and how to analyze it.”