Chapter 1 Data Science

The definition of data science can be a little slippery. One current view of data science, is exemplified by Steven Geringer’s 2014 Venn diagram.

Data Science Venn Diagram from Steven Geringer

Figure 1.1: Data Science Venn Diagram from Steven Geringer

  • The field encompasses ideas from mathematics and statistics and from computer science, but with a heavy reliance on subject-matter knowledge. In our case, this includes clinical, health-related, medical or biological knowledge.
  • As Gelman and Nolan (2017) suggest, the experience and intuition necessary for good statistical practice are hard to obtain, and teaching data science provides an excellent opportunity to reinforce statistical thinking skills across the full cycle of a data analysis project.
  • The principal form in which computer science (coding/programming) play a role in this course is to provide a form of communication. You’ll need to learn how to express your ideas not just orally and in writing, but also through your code.

1.1 Why a unicorn?

Data Science is a team activity. Everyone working in data science brings some part of the necessary skillset, but no one person can cover all three areas alone for excellent projects.

[The individual who is truly expert in all three key areas (mathematics/statistics, computer science and subject-matter knowledge) is] a mythical beast with magical powers who’s rumored to exist but is never actually seen in the wild.

1.2 Data Science Project Cycle

A typical data science project can be modeled as follows, which comes from the introduction to the amazing book R for Data Science, by Garrett Grolemund and Hadley Wickham, which is a key text for this course (Grolemund and Wickham 2017).

Source: R for Data Science: Introduction

Figure 1.2: Source: R for Data Science: Introduction

This diagram is sometimes referred to as the Krebs Cycle of Data Science. For more on the steps of a data science project, we encourage you to read the Introduction of Grolemund and Wickham (2017).

1.3 What Will We Discuss in 431?

We’ll discuss each of these elements in the 431 course, focusing at the start on understanding our data through transformation, modeling and (especially in the early stages) visualization. In 431, we learn how to get things done.

  • We get people working with R and R Studio and R Markdown, even if they are completely new to coding. A gentle introduction is provided at Ismay and Kim (2017)
  • We learn how to use the tidyverse(, an array of tools in R (mostly developed by Hadley Wickham and his colleagues at R Studio) which share an underlying philosophy to make data science faster, easier, more reproducible and more fun. A critical text for understanding the tidyverse is Grolemund and Wickham (2017). Tidyverse tools facilitate:
    • importing data into R, which can be the source of intense pain for some things, but is really quite easy 95% of the time with the right tool.
    • tidying data, that is, storing it in a format that includes one row per observation and one column per variable. This is harder, and more important, than you might think.
    • transforming data, perhaps by identifying specific subgroups of interest, creating new variables based on existing ones, or calculating summaries.
    • visualizing data to generate actual knowledge and identify questions about the data – this is an area where R really shines, and we’ll start with it in class.
    • modeling data, taking the approach that modeling is complementary to visualization, and allows us to answer questions that visualization helps us identify.
    • and last, but definitely not least, communicating results, models and visualizations to others, in a way that is reproducible and effective.
  • Some programming/coding is an inevitable requirement to accomplish all of these aims. If you are leery of coding, you’ll need to get past that, with the help of this course and our stellar teaching assistants. Getting started is always the most challenging part, but our experience is that most of the pain of developing these new skills evaporates by early October.
  • Having completed some fundamental work in Part A of the course, we then learn how to use a variety of R packages and statistical methods to accomplish specific inferential tasks (in Part B, mostly) and modeling tasks (in Part C, mostly.)


Gelman, Andrew, and Deborah Nolan. 2017. Teaching Statistics: A Bag of Tricks. Second. Oxford, UK: Oxford University Press.

Grolemund, Garrett, and Hadley Wickham. 2017. R for Data Science. O’Reilly.

Ismay, Chester, and Albert Y. Kim. 2017. ModernDive: An Introduction to Statistical and Data Sciences via R.

You are watching: Data Science for Biological, Medical and Health Research: Notes for 431. Info created by GBee English Center selection and synthesis along with other related topics.