Eric Weber doesn't begin teaching data science with code or algorithms. Instead, the Iowa State University mathematics professor asks his future teachers to imagine the scientific method running backward.
In traditional science, you start with a hypothesis and test it. Data science flips that script entirely. You begin with data that already exists—numbers collected years ago, information gathered for a completely different purpose—and work backward to discover patterns, connections, and surprises. Those clues then help you figure out what questions you should even be asking.
This insight, grounded in concrete pedagogy rather than technical jargon, underpins a curriculum that Weber and colleagues at Iowa State and the University of Northern Iowa have developed to prepare future math teachers to teach data science in high-school classrooms. The collaboration reflects a growing national push to embed data science literacy into secondary education. Multiple professional societies in mathematics, statistics, and mathematics education have already released statements supporting data-science courses in high schools—yet teachers expected to lead these courses often receive little to no preparation.
Weber, along with co-authors Heather Gallivan (associate professor of mathematics education at UNI), Lydia Butters (a former UNI math education student now teaching at Cedar Falls High School), and Stephen Nathan Mercil (a former Iowa State mathematics doctoral student now instructor at the University of St. Thomas, Minnesota), designed their five-week module to close that gap by building on what pre-service teachers already know. A regression line becomes a model. A classification problem becomes a geometry puzzle. An optimization routine becomes a function-minimizing exercise. By anchoring data science in familiar mathematical structures from algebra, geometry, and calculus, the curriculum sidesteps the intimidation factor that often stalls new learners.
"If we can break down the initial barrier of, 'I don't know what data science is,' then their ability to make that transition becomes pretty quick," Weber explained.
The project began in 2019 when Weber and Mercil first piloted the curriculum at Iowa State. The first full run happened in spring 2020—just as the pandemic forced classes online. After Weber partnered with Gallivan, whose background in statistics enriched the universities' combined approach, funding allowed the team to refine the lessons and roll out the curriculum at both campuses starting in 2023. Since then, the module has been taught every spring, with improvements added each year based on student feedback and classroom experience.
To ground the curriculum in real-world application, Weber's team uses a mix of synthetic and authentic datasets. One dataset simulates animal-tracking information—timestamps, locations, and headings—giving students hands-on practice with visualization, dimensionality reduction, and prediction. Another draws on housing data collected by local high-school students, allowing pre-service teachers to practice multiple regression and envision how they might guide their own students through similar projects. These examples help future teachers understand that data science isn't some separate universe from the mathematics they already study. It's built directly on it.
Weber, who now serves on a committee assembled by the Iowa Department of Education to help write data science learning standards for the state, sees this work as part of a larger shift. As schools across the country add data-science courses, preparing teachers becomes not just important but urgent. And starting with future educators grounded in mathematics provides the strongest foundation yet.