Week 3, Part 1: Data Wrangling with dplyr and Data Ethics

1 Learning Objectives

  • Use the five main dplyr verbs:

    • filter()
    • arrange()
    • select()
    • mutate()
    • summarize()

  • Use group_by() to perform groupwise operations

  • Use the pipe operator (|>) to chain together data wrangling operations

📖 Readings: 45-60 min

💻 Optional Activity: 30 min

📽 Optional Videos: 60 min

2 Data Structures – list, data.frame, and tibble

We had a brief introduction to data structures in R on the first day, but it is useful to re-visit some of these concepts now and introduce a new structure called a tibble.

list and data.frame

We are exclusively working with “rectangular” data in this class - where there are n rows, and each column represents a variable. These columns can be of multiple different data types (i.e. string, boolean, integer, date, etc.), so we can’t use a matrix to represent our data typically. This is because matrices are homogeneous data structures, meaning all values need to be of one data type.

As we learned earlier in the course, a list is a one-dimensional column of heterogeneous data – i.e. the things stored in a list can be of different types. For example, I could create a list like this:

mylist <- list(x = c(1, 2, 3),
               y = 3,
               d = matrix(c("a", "b", "c", "d"), ncol = 2))

Lists are a super powerful data structure and offer a lot of flexibility.

When working with rectangular data, we don’t need (or want) quite as much flexibility as a pure list. Since rectangular data is so common and we more or less know the kinds of things we want to do with it, there is a special data structure in R called a data.frame. A data.frame is really a special kind of list, where every element of is a column of equal length. Many functions in R are built to work with data.frames, including all of the dplyr functions that we are learning with this week.

tibble

Essentially, tibbles are “fancy” data frames. They operate the same as data frames, with some extra features that can make them easier to work in certain circumstances. You may notice that when you print a table after applying dplyr functions to it, it will be identified as a tibble. For the purposes of this class, the distinctions between data.frames and tibbles are only important when we start doing more complex analyses towards the end of the course. If you are curious now, you can read more about tibbles on the old R4DS.

3 Data Wrangling with dplyr

📖 Required Reading: R4DS Data Transform

📽 Optional Videos on dplyr verbs

💻 Optional Tutorial for more practice: Practice with dplyr

4 Data Ethics

📖 Required Reading: The Numbers Don’t Speak for Themselves

Some things to think about while you read to prepare for our class discussion:

  • What was the main take-away for you?
  • What points stood out to you?
  • What was something that suprised you?
  • Is there anything you didn’t agree with?
  • What questions do you have after reading?