Introductory Data Science Courses

Introduction

Data science, as a field, is establishing itself but data science education is still very much evolving. Degree programs in data science have proliferated rapidly over the past decade, and the scholarship of teaching and learning that supports them is being built alongside them.

A field of flavors

The current landscape of Introductory Data Science courses is, in part, a period of experimentation: institutions are testing different entry points, different balances of computation and inference, and different sequencing of topics. Some of this variation will narrow as the field matures.

But we expect a good deal of this variation to persist. Data science sits at the intersection of statistics, computer science, mathematics, and a wide range of application domains, and departments approach that intersection from different directions depending on their disciplinary home, their students, and their institutional mission.

The result is not a single course waiting to be discovered, but a family of legitimate “flavors” of introductory data science; courses that will continue to differ in emphasis even after the field’s early uncertainty settles.

Two shaping realities

This experimentation is further complicated by two realities.

  1. An introductory course often stands in for the discipline itself. For many students, this course is their exposure to data science’s way of thinking: its questions, its habits of mind, its methods. The department offering it becomes, in effect, a proxy for the field as a whole.

  2. Introductory data science courses are not a single, uniform offering. They serve a range of purposes and serve students with a range of preparation:

    • The required first step into a data science major
    • A general education requirement
    • A course designed primarily to build broad data literacy for students who will take no further coursework in the area

    These courses differ accordingly in prerequisites, pacing, and depth.

The absence of a settled framework is not a reason to defer good practice. It is precisely because the field of data science education is still defining itself that shared recommendations are needed now, to give this period of experimentation some common direction.

What this means for these guidelines

Despite this variation, we believe strongly that the ten recommendations offered in these guidelines are essential, and that they should be reflected in the curriculum of any introductory data science course regardless of its specific purpose, audience, or institutional context.

Introductory statistics education benefited from decades of research, pedagogical practice, and prior GAISE guidelines that allowed this GAISE College Report revision to converge on a set of learning outcomes; introductory data science education has no comparable body of accumulated practice to draw on yet. There is not yet an accepted framework, no settled agreement on the knowledge, skills, and abilities undergraduate students should acquire in a data science program, let alone in a single introductory course. That is why this report offers recommendations rather than a settled set of SLOs for introductory data science at this time.

New research: See what introductory data science courses actually look like.

  • What is being taught
  • Who is teaching it
  • Where it’s taught

Explore The Landscape of Introductory Data Science Courses →