A Fresh Look at Introductory Data Science

Figures & data

Table 1 Summary of programming languages used in each course and the estimated breakdown of percent of class time spent on various course components.

	Duke	Berkeley	Cambridge	Smith	Stanford
Programming	R	Python	Pseudocode	R, SQL	R
language
Data visualization	15%	5%	0%	32%	10%
Data wrangling	10%	15%	0%	36%	0%
Other EDA	10%	5%	0%	12%	10%
Inference	20%	30%	25%	0%	50%
Modeling	25%	20%	35%	0%	20%
Programming	10%	10%	0%	5%	0%
principles
Mathematical	5%	5%	35%	0%	0%
foundations/theory
Communication	5%	5%	0%	10%	10%
Ethics	0%	5%	5%	5%	0%

Fig. 1 Flow of topics in Introduction to Data Science and Statistical Thinking at Duke University.

Table 2 Topics previously covered in Unit 3 of the course.

Topic	Synopsis	Duration
Data science ethics	Misrepresentation of results in data visualizations and reporting, data privacy and data breaches, gender bias in machine translated text, algorithmic bias and race in sentencing and parole length decisions.	1–2 class periods
Interactive reporting and visualization with Shiny	Introduce the basics of the shiny package for building interactive web applications and build a simple application for browsing data on movies.	1 class period
Building static dashboards	Build static dashboards using the flexdashboard package.	1 class period
Building interactive dashboards	Build interactive dashboards using the shiny and flexdashboard packages.	2 class periods
Text mining	Perform basic text mining techniques (e.g., sentiment analysis, term frequency-inverse document frequency) using the tidytext package and data on song lyrics (retrieved with the genius package) or on books (retrieved with the gutenbergr package).	1 class period
Bayesian inference	Introduction to Bayesian inference as a way of decision making using data on sensitivity and specificity of breast cancer screening tests.	1 class period

Fig. 2 Relationship between number of commits and final course grade for each student at three time points in the semester.