Tufts to launch new graduate initiative designed to connect data science, policy

Colorful trees are pictured near Tisch Library on Oct. 22. Nicole Garay / The Tufts Daily

Tufts will launch a new program called “[email protected]: Educating Policy-Savvy Data Experts and Data-Proficient Decision Makers.” This interdisciplinary initiative aims to bridge gaps between science, technology, engineering and math (STEM) and non-STEM disciplines to better inform policymaking.

The National Science Foundation (NSF) is funding the program through a grant amounting to almost $3 million; it was awarded under the NSF’s research traineeship program. Led by Shafiqul Islam, a professor of civil and environmental engineering, the project will include nine faculty members and over 140 graduate students from across disciplines and schools at Tufts, according to Islam.

“This is about re-inventing … graduate education in the data sciences,” Abani Patra, director of the Data Intensive Studies Center at Tufts and one of the co-principal investigators of the grant, said. 

Patra believes the grant will be instrumental in building Tufts’ data science research and education programs.

“The grant allows you to explore new ideas [and] come up with new paradigms that then have to be institutionalized and sustained,Patra said. “One of the goals [of the grant] is raising institutional capacity in data science education and training”. 

Patra further explained that [email protected] will equip graduate students to analyze data in a specific context.

“It’s one thing to learn about data in the abstract — ‘here’s some data to go play with.’ It’s quite another thing to learn about data in context,” he said. 

In their proposal to the NSF, Islam and his team detailed the disconnect between those with a deep background in data science and those with more policy-oriented expertise.

“The failure to bridge the gaps between numbers and narratives creates significant bottlenecks to evidence-based, interdisciplinary decision making, especially at the ‘messy’ interface between human and natural systems,” the proposal reads.

David Hammer, a co-principal investigator, chair of the Department of Education and a professor of education and physics and astronomy, echoed the statement.

There are the engineers who can calculate things but can’t construct compelling narratives about what they’re doing, and then the other side is the social science, humanities sorts of people … [the] kind of folks who can construct narratives but don’t have a sense of the concepts in data science,” Hammer said. “And so the idea of this is to build up something that brings those together.”

Islam noted that the program intends to train students into two types of professionals: “policy-savvy data experts” for those with a STEM background to gain ground on policy analysis, as well as “data-savvy decision makers” for those from non-STEM disciplines to acquire rigorous quantitative skills.

Graduate students at Tufts will be able to enroll in [email protected] concurrently with their graduate degree program, according to Hammer. He further explained that the program will be weaved into courses as “modular course elements,” which will allow students to complete the program without lengthening the time it takes to graduate.

Islam elaborated on the four integrated data skill areas that make up the modular course elements.

“If you want to be a data scientist, you need some formal knowledge, you may need some practice techniques, you need some thoughtful practice and you need some wisdom,” Islam said.

Hammer also shared that in addition to these elements, the program includes another component named “problem-focused immersion.”

Islam explained that the initiative will invite its students to meet every other week to discuss a particular problem, and they will subsequently create a project to address the problem. A student who completes this component will receive a certificate.

Hammer illustrated how problem-focused immersion would function, using COVID-19 as an example.

“Analyzing and making recommendations for [COVID-19] policy could be an example of such a problem which is an area for data science … what is affecting the rate of infection, how much of it is urban density, how much of it is mask-wearing, how much of it is local policies,” Hammer said. “[That could become] a problem-based immersion for a team of people with complementary expertise to be working on.”

Islam said he expects the program to start fall 2021.