Seated in small groups with their laptops, students from lecturer Victoria Robinson’s introductory Ethnic Studies course are talking excitedly while poring over data from private immigration detention centers across California. Their work together is part of a data-centric project they will present at the end of the term.
Robinson has taught the course for several years, leading students to examine the connections between immigration enforcement and criminal justice systems. But this spring marks the first time that a group of Robinson’s students are using data science as a lens through which to address topics of incarceration and immigration.
“They’re asking, ‘what do the data tell me?’” said Robinson, also the program director of the American Cultures program. “The data science tools actually fit in great sympathy with the other tools of critical analysis that we teach in Ethnic Studies.”
Students in Victoria Robinson's Ethnic Studies course working together on a data science module project
Partnerships between faculty and data science specialists
Robinson is one of a number of faculty who are now incorporating “data science modules” into their courses in different fields across Berkeley. Modules are short explorations into data science, in which faculty offer students the opportunity to work hands-on with a dataset relevant to their course, and some instruction on the principles of statistics and computing.
The modules are created through collaboration between faculty and experienced student teams who are part of the Data Science Education Program. The modules teams, which include undergraduate and graduate students as well as D-Lab instructional staff and Berkeley Institute for Data Science (BIDS) fellows, meet with faculty to help them build a lesson for a current or future class, tailoring the module to the faculty’s teaching and research objectives.
Faculty say the modules introduce a new perspective and methodology for teaching and research within their domain of expertise. Students say that modules enable them to go beyond reading and critiquing research done by others; instead, they can pose their own questions and undertake their own analysis using data.
“For faculty, it’s about empowering their students to do their own research and go further in their fields,” said Alexander Ivanoff, a junior studying Cognitive Science and Statistics and the student leader of the modules team. “For students, modules really take away the intimidation factor by building the data science toolsets into subjects they’re already interested in.”
Using high-tech tools
As in Data Science Education Program courses, students in the modules learn to use the Python programming language and complete assignments in Jupyter notebooks, both gold-standard data science tools. Faculty can share data with students, assign homework and write instructions in the cloud-based notebook (no software installation required), and then ask for students’ interpretation of the results, all within the same tool.
Applying data science to diverse fields
The modules team is available to work with faculty to use the tools and design the module to fit the course objectives. Data science modules are as diverse as the myriad disciplines taught across Berkeley. Below are a few examples of the ways that Berkeley faculty are adding data science modules to their courses.
In Niek Veldhuis’ Near Eastern Studies 101B course, students analyzed data from ancient Sumerian texts to plot storylines and understand characters.
In Amy Tick’s Rhetoric 1A course, students performed their own research analyses on the demographics that influence voting patterns, using current polling and demographic data during the last election.
In Laura Nelson’s Gender and Women’s Studies course, a module included a lecture on the use of statistics in the news media, and how students can question and understand statistics in news articles.
In Neil Davies’ Ecology course, students practice field ecology in a hands-on setting in the island of Moorea, with a new module focused on data management best practices, population modeling and other useful tools for ecologists.
Students learning from each other
Julian Kudszus, a junior Computer Science major and lead for Victoria Robinson’s Ethnic Studies module, said students in the course range from Computer Science majors to humanities majors, who work collaboratively together on projects. The collaboration among students of various majors and technical backgrounds makes for a rich learning environment, Kudszus said. Peer-to-peer learning, a common dynamic in Computer Science courses, is a big part of the learning experience in data science modules.
“This semester we’ve heard from several students without programming experience who said, ‘Hey, this is something I can do!”, said Kudszus. “The modules do a good job of breaking down the barriers to entry and help students realize that data science is very accessible.”
Seven modules are currently underway or in various stages of development, with interest ramping up significantly over the past several months as more faculty learn about the modules offerings.
Looking ahead, the modules team is now developing standardized Jupyter notebooks that can be starting points for specific domains, ranging from the biological sciences to social sciences to engineering. The new Jupyter notebooks will enable faculty to find domain-specific resources -- including data, assignments, tools and more -- so that they can explore how to add data science modules into their courses. For example, one such Jupyter notebook will highlight ways for instructors to use text analysis, a common analytical technique that is increasingly used in history, literature, linguistics, and more.
In addition, the team is looking to build greater connections and partnerships with graduate students as well with other organizations that support data science, including the D-Lab.