A Fresh Look at Introductory Data Science

Abstract

The proliferation of vast quantities of available datasets that are large and complex in nature has challenged universities to keep up with the demand for graduates trained in both the statistical and the computational set of skills required to effectively plan, acquire, manage, analyze, and communicate the findings of such data. To keep up with this demand, attracting students early on to data science as well as providing them a solid foray into the field becomes increasingly important. We present a case study of an introductory undergraduate course in data science that is designed to address these needs. Offered at Duke University, this course has no prerequisites and serves a wide audience of aspiring statistics and data science majors as well as humanities, social sciences, and natural sciences students. We discuss the unique set of challenges posed by offering such a course, and in light of these challenges, we present a detailed discussion into the pedagogical design elements, content, structure, computational infrastructure, and the assessment methodology of the course. We also offer a repository containing all teaching materials that are opensource, along with supplementary materials and the R code for reproducing the figures found in the article.

Computational skills by stealth in introductory data science teaching

“…With extremely high‐level code like this, where small pieces of code deliver rich information, students' first experiences of direct interactions with code can associate it with power, not drudgery (they get rich graphics and information with very little effort). This is very similar to the philosophy espoused by statisticians and statistics educators over the past two decades [5], of the importance of the first‐day hook: get students up and running with powerful, interactive data analysis as early as functionally possible, so that the power overwhelms the fear.…”

Teaching Creative and Practical Data Science at Scale

“…We are instructors involved in designing and teaching new, large courses on data science that aim to teach hands-on data analysis through computational work. Building on the lessons learned from having run these courses for several thousand undergraduate students and on the work of other data science educators (Baumer 2015;Carver et al 2016;Çetinkaya-Rundel and Ellison 2020;Hicks and Irizarry 2018;Loy et al 2019; National Academies of Sciences, Engineering and Medicine 2018; Wood et al 2018;Yan and Davis 2019), in this article, we discuss our approach to, and goals for, teaching data science. In these goals, we find much alignment with the vision of Nolan and Temple Lang, including their positioning of statistics (or data analysis more broadly) as flexible problem solving, and in their call for pedagogy of data analytical skills to be grounded in practical work with real datasets, to which we also add a goal of scaling data education to meet demand.…”

“…Within projects, students must have opportunities to write good and well-documented code that accomplishes their analysis goals (Wilson et al 2014(Wilson et al , 2017. Throughout this entire process, version control should be used to facilitate collaboration with other students, code sharing, and portfolio generation (Blischak et al 2016;Çetinkaya-Rundel and Ellison 2020).…”

infer: An R package for tidyverse-friendly statistical inference

“…Such a grammar lends itself to applications in teaching, data pedagogy research, applied scientific research, and advanced predictive modeling. For one, the principled approach of the infer package has made it an especially good fit for teaching introductory statistics and data science (Baumer et al, 2020;Çetinkaya-Rundel & Ellison, 2021;Ismay & Kim, 2019) and research in data pedagogy (Fergusson & Pfannkuch, 2021;Loy, 2021). Further, the package has already seen usage in a number of published scientific applications (Ask et al, 2021;Fallon & Hinds, 2021;McLean et al, 2021).…”