pedagogy 2019
June 16, 2019

Data science is an intellectual force multiplier. It can unlock insights on social problems, art collections, and language. It’s the glue that brings disciplines together. 

These are among the threads that drew three dozen Berkeley faculty and lecturers from across campus earlier this month for a three-day “boot camp” on incorporating data science into their teaching. They came from African studies, economics, law, and libraries; from statistics, history, and public health.

Participants said they not only wanted to empower themselves and their students with data science tools and open platforms; they wanted to cultivate better understanding of how to use data responsibly—to invest students with the ability to see and avoid practices that can augment inequities, for instance, or amplify misinformation.

At the fourth annual Division of Data Sciences Pedagogy Workshop, they covered all that and more. The workshop, like Berkeley’s Foundations of Data Science (Data 8) class, is designed to meet participants where they are, both in terms of interest and ability; to make the fundamentals of data science broadly accessible and relevant.

As workshop organizer Eric Van Dusen put it; “I’m here to work with you and serve you, and to get you excited about the possibilities.” For many faculty, the next step is to collaborate with Eric and his team creating data science modules, data-centric curriculum or projects for existing classes.

Exploring Data in Context

The Division of Data Sciences’ Data Science Modules program is powered by a team of students who work to create short explorations into data science by designing data sets in cloud-based Jupyter (link is external)notebooks that enable students to explore topics relevant to their course, whether it’s economics or ethnic studies.

Using these datasets, they design mini coding lessons, charts and graphs, allowing students to experience how data can be used to better understand the course content.

Karla Palos Castellanos, a team lead in the data science Modules program, originally became interested after taking a module herself—Social Inequalities, American Cultures (Sociology 130AC). She was intrigued when she discovered similar explorations were being created for courses across campus.  

Since joining the student team, she has worked on modules in public health, engineering, ethnic studies, and sports analytics. Before joining the modules team, Karla worked as a lab assistant, and she saw designing modules as a different way of teaching. One of the most important stages of designing a module, she said, is the planning phase. The goal during this portion is understanding the strengths and weaknesses of the audience—the students enrolled in this particular class—as well as understanding the teacher’s learning goals. For some modules, because of the nature of the class, professors are more interested in having students learn to actually create graphs, while in other cases, the focus may be more on interpreting them.

“It’s a lot of putting yourself in [the student’s] position and knowing how much they know,” Karla said. “[Thinking about] how you felt when you were a beginning programmer and what things you wish someone had walked you through.”

From Urban Studies to French to Engineering

Keeley Takimoto, who originally taught the sociology module that piqued Palos’ interest, was part of the data science modules team when she was a Berkeley student. After graduating, she took on a job as the Modules Curriculum Lead within the Division of Data Sciences and is now in charge of overseeing the student teams working on a variety of different modules.

Her job involves working directly with professors to understand the goals of their module and their class. Last semester, she said, there were students working on courses ranging from urban data analysis to French culture to engaged engineering.

Recently, the modules team has also focused on developing more classes in collaboration with American Cultures. These courses are particularly appealing, Keeley said, because they are large, diverse courses with students who are not necessarily expecting computing.

More than 1,700 students across 50 courses have taken a data science modules class, and interest is expected to grow in the upcoming years as more instructors, sometimes at their students’ urging, see the potential for incorporating data science to drive insights and discovery in their courses. Every new pedagogy workshop generates new ideas for modules; several prospects are already percolating in the wake of the workshop earlier this month.

In addition to the summer workshop, the Division offers similar but shorter programs for Berkeley faculty throughout the year. And next week, more than 60 educators from institutions across the country will come to Berkeley to learn and share ideas about teaching data science at the 2019 National Workshop on Data Science Education.