What’s it all about, Alfie?

Lecture 24

John Zito

Duke University
STA 199 Spring 2025

2025-04-22

Course admin

While you wait

• Go to your ae project in RStudio.

• Make sure all of your changes up to this point are committed and pushed, i.e., there’s nothing left in your Git pane.

• Click Pull to get today’s application exercise file: ae-20-final-review.qmd.

• Wait until you’re prompted to work on the application exercise during class before editing the file.

Projects due tomorrow at 5PM

• Written report in the index file;
• Code chunks should not be displayed;
• Five minute video presentation;
• It needs to be obvious that all team members participated.

Final exam next Tuesday 4/29 at 9AM

• Roughly the same length as a midterm;
• You have three hours to complete it;
• You get both sides of an 8.5” x 11” sheet of notes;
• All multiple choice;
• Room assignments emailed to you;
• 20% of your final course grade;
• Replaces a lower in-class midterm score;
• Visit my OHs to view old midterm solutions.

Style of questions

Things you’ve seen that will reappear:

• Do you understand this picture?
• Can you debug this incorrect code?
• What is the correct statistical interpretation of this quantity?

A skill you will need:

Can you picture the intermediate output of a data pipeline.

From Midterm 1…

# A tibble: 6 × 3
  age   opinion                n_people
  <chr> <chr>                     <dbl>
1 18-49 Remain Against The Law       59
2 18-49 Be Made Legal               292
3 18-49 Not Sure                     40
4 50+   Remain Against The Law       67
5 50+   Be Made Legal               245
6 50+   Not Sure                     68

From Midterm 1…

Why does this give an error?

survey_counts |>
  group_by(age) |>
  summarise(
    age_sum = sum(n_people)
  ) |> 
  mutate(
    prop = n_people / sum(n_people)
  )

Error in `mutate()`:
ℹ In argument: `prop = n_people/sum(n_people)`.
Caused by error:
! object 'n_people' not found

How do the intermediate steps look?

survey_counts 

# A tibble: 6 × 3
  age   opinion                n_people
  <chr> <chr>                     <dbl>
1 18-49 Remain Against The Law       59
2 18-49 Be Made Legal               292
3 18-49 Not Sure                     40
4 50+   Remain Against The Law       67
5 50+   Be Made Legal               245
6 50+   Not Sure                     68

How do the intermediate steps look?

survey_counts |>
  group_by(age)

# A tibble: 6 × 3
# Groups:   age [2]
  age   opinion                n_people
  <chr> <chr>                     <dbl>
1 18-49 Remain Against The Law       59
2 18-49 Be Made Legal               292
3 18-49 Not Sure                     40
4 50+   Remain Against The Law       67
5 50+   Be Made Legal               245
6 50+   Not Sure                     68

How do the intermediate steps look?

survey_counts |>
  group_by(age) |>
  summarise(
    age_sum = sum(n_people)
  )

# A tibble: 2 × 2
  age   age_sum
  <chr>   <dbl>
1 18-49     391
2 50+       380

Indeed, n_people is not a column in the data frame that is being piped into mutate, hence the error.

How do the intermediate steps look?

survey_counts |>
  group_by(age) |>
  summarise(
    age_sum = sum(n_people)
  ) |> 
  mutate(
    prop = n_people / sum(n_people)
  )

Error in `mutate()`:
ℹ In argument: `prop = n_people/sum(n_people)`.
Caused by error:
! object 'n_people' not found

The full monty

• Preparation: importing, joining, reshaping, mutating, etc;

• Exploration: pictures and summaries;

• Modeling: linear/logistic regression, model selection, etc.

• Inference: interval estimation and hypothesis testing.

ae-20-final-review

• Go to your ae project in RStudio.

• If you haven’t yet done so, make sure all of your changes up to this point are committed and pushed, i.e., there’s nothing left in your Git pane.

• If you haven’t yet done so, click Pull to get today’s application exercise file: ae-20-final-review.qmd.

• Work through the application exercise in class, and render, commit, and push your edits.