Describing data: part 4

Lecture 10

Aidan Combs

Duke University
SOCIOL 333 - Summer Term 1 2023

2023-06-05

Logistics

• Project component 2: descriptive statistics

  • All materials are posted (instructions, example, github repos)
  • Due Wednesday June 7 11:59pm
  • I have returned feedback (GitHub issue on your repo) and grades (Sakai) for your proposals
  • We have covered everything for parts 1-3. Part 4 is today and tomorrow.
  • I’m around before/after class for questions (after generally better), or set up an appointment

Today

• Answers to filtering
• Pipes
• Working on creating variables
• Answers to creating variables
• Plots

Filtering exercise (ex-5-31): answers!

• Find this on your computer (no need to clone it again)

Stringing commands together with pipes (|>)

• Often we need to change data frames in more than one way
• Example from last week: How does employment status vary by age category?
• We need to create an age category variable (as we talked about Thursday)
• But we probably want to filter too–why analyze kids?

Stringing commands together: approach 1

We can do this with the tools we have.

# first we make the new variable
acs12_newagevar <- mutate(acs12, agecat = case_when(age < 14 ~ "child",
                                                    age < 18 ~ "teen",
                                                    age < 67 ~ "adult",
                                                    TRUE ~ "retired"))

# then we filter to remove children
acs12_nokids <- filter(acs12_newagevar, agecat != "child")

# did it work?
table(acs12_nokids$agecat, useNA = "always")

  adult retired    teen    <NA> 
   1264     297      89       0 

• But it’s kind of ugly… we don’t need to save that middle step

Stringing commands together: approach 2, with |>

• The pipe operator, |>, lets us pass the result of one function directly into another one
• The | symbol is the key below “delete” on your keyboard (not I, not l, not 1)
• It replaces the first mutate/filter argument (the dataset)
• “Take the thing that came before this and give it to the function that comes after this”

acs12_nokids <- acs12 |> # start with acs12
  mutate(agecat = case_when(age < 14 ~ "child", #then add a new variable to it
                            age < 18 ~ "teen",
                            age < 67 ~ "adult",
                            TRUE ~ "retired")) |> 
  filter(agecat != "child") # then filter out the kids

table(acs12_nokids$agecat, useNA = "always")

  adult retired    teen    <NA> 
   1264     297      89       0 

Another |> example

• This:

mutate(acs12, agecat = case_when(age < 14 ~ "child",
                                 age < 18 ~ "teen",
                                 age < 67 ~ "adult",
                                 TRUE ~ "retired"))

• Is the same as this:

acs12 |> 
  mutate(agecat = case_when(age < 14 ~ "child",
                            age < 18 ~ "teen",
                            age < 67 ~ "adult",
                            TRUE ~ "retired"))

More on |>

• We’re just scratching the surface in this class

• When you need to clean your data or when your analyses are more complex, |> makes your life a lot easier!

• Sometimes in internet resources more than ~2 years old, you’ll see %>% instead–this older version does exactly the same thing

• Annoying to type out? There’s a keyboard shortcut

  • on mac, command-shift-M
  • on pc, ctrl-shift-M

Continuing on creating new variables

• Find your exercise from last Thursday (ex-6-01)

  • No need to clone it again; find where it’s saved on your computer

• There are hints on the website–use them if you get stuck

• 10-15 minutes to work

Creating new variables: answers!