Exam 1 review

Lecture 9

Dr. Mine Çetinkaya-Rundel

Duke University
STA 199 - Spring 2024

2024-02-13

Warm up

While you wait for class to begin…

Questions for/about the exam?

Packages

library(tidyverse)
library(openintro)

Announcements

• Exam format / flow

• Academic dishonesty / Duke Community Standard

Type coercion

Explicit vs. implicit type coercion

• Explicit type coercion: You ask R to change the type of a variable

• Implicit type coercion: R changes / makes assumptions for you about the type of a variable without you asking for it

  • This happens because in a vector, you can’t have multiple types of values

Vectors

• A vector is a collection of values

  • Atomic vectors can only contain values of the same type

  • Lists can contain values of different types

• Why do we care? Because each column of a data frame is a vector.

df <- tibble(
  x = c(1, 2, 3),          # numeric (double)
  y = c("a", "b", "c"),    # character
  z = c(TRUE, FALSE, TRUE) # logical
)
df

# A tibble: 3 × 3
      x y     z    
  <dbl> <chr> <lgl>
1     1 a     TRUE 
2     2 b     FALSE
3     3 c     TRUE 

Explicit coercion

✅ From numeric to character

df |>
  mutate(x_new = as.character(x))

# A tibble: 3 × 4
      x y     z     x_new
  <dbl> <chr> <lgl> <chr>
1     1 a     TRUE  1    
2     2 b     FALSE 2    
3     3 c     TRUE  3    

Explicit coercion

❌ From character to numeric

df |>
  mutate(y_new = as.numeric(y))

# A tibble: 3 × 4
      x y     z     y_new
  <dbl> <chr> <lgl> <dbl>
1     1 a     TRUE     NA
2     2 b     FALSE    NA
3     3 c     TRUE     NA

Implicit coercion

Which of the column types were implicitly coerced?

df <- tibble(
  w = c(1, 2, 3),
  x = c("a", "b", 4),
  y = c("c", "d", NA),
  z = c(5, 6, NA),
)
df

# A tibble: 3 × 4
      w x     y         z
  <dbl> <chr> <chr> <dbl>
1     1 a     c         5
2     2 b     d         6
3     3 4     <NA>     NA

Collecting data

Suppose you conduct a survey and ask students their student ID number and number of credits they’re taking this semester. What is the type of each variable?

survey_raw <- tibble(
  student_id = c(273674, 298765, 287129, "I don't remember"),
  n_credits = c(4, 4.5, "I'm not sure yet", "2 - underloading")
)
survey_raw

# A tibble: 4 × 2
  student_id       n_credits       
  <chr>            <chr>           
1 273674           4               
2 298765           4.5             
3 287129           I'm not sure yet
4 I don't remember 2 - underloading

Cleaning data

survey <- survey_raw |>
  mutate(
    student_id = if_else(student_id == "I don't remember", NA, student_id),
    n_credits = case_when(
      n_credits == "I'm not sure yet" ~ NA,
      n_credits == "2 - underloading" ~ "2",
      .default = n_credits
    ),
    n_credits = as.numeric(n_credits)
  )
survey

# A tibble: 4 × 2
  student_id n_credits
  <chr>          <dbl>
1 273674           4  
2 298765           4.5
3 287129          NA  
4 <NA>             2  

Cleaning data – alternative

survey <- survey_raw |>
  mutate(
    student_id = parse_number(student_id),
    n_credits = parse_number(n_credits)
  )
survey

# A tibble: 4 × 2
  student_id n_credits
       <dbl>     <dbl>
1     273674       4  
2     298765       4.5
3     287129      NA  
4         NA       2  

Recap: Type coercion

• If variables in a data frame have multiple types of values, R will coerce them into a single type, which may or may not be what you want.

• If what R does by default is not what you want, you can use explicit coercion functions like as.numeric(), as.character(), etc. to turn them into the types you want them to be, which will generally also involve cleaning up the features of the data that caused the unwanted implicit coercion in the first place.

Aesthetic mappings

openintro::loan50

loan50 |>
  select(annual_income, interest_rate, homeownership)

# A tibble: 50 × 3
   annual_income interest_rate homeownership
           <dbl>         <dbl> <fct>        
 1         59000         10.9  rent         
 2         60000          9.92 rent         
 3         75000         26.3  mortgage     
 4         75000          9.92 rent         
 5        254000          9.43 mortgage     
 6         67000          9.92 mortgage     
 7         28800         17.1  rent         
 8         80000          6.08 mortgage     
 9         34000          7.97 rent         
10         80000         12.6  mortgage     
# ℹ 40 more rows

Aesthetic mappings

What will the following code result in?

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate, color = homeownership, shape = homeownership)
) +
  geom_point() +
  scale_color_colorblind()

Aesthetic mappings

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate, color = homeownership, shape = homeownership)
) +
  geom_point() +
  scale_color_colorblind()

Global mappings

What will the following code result in?

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate, color = homeownership, shape = homeownership)
) +
  geom_point() +
  geom_smooth(se = FALSE) +
  scale_color_colorblind()

Global mappings

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate, color = homeownership, shape = homeownership)
) +
  geom_point() +
  geom_smooth(se = FALSE) +
  scale_color_colorblind()

Local mappings

What will the following code result in?

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate)
) +
  geom_point(aes(color = homeownership)) +
  geom_smooth(se = FALSE) +
  scale_color_colorblind()

Local mappings

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate)
) +
  geom_point(aes(color = homeownership)) +
  geom_smooth(se = FALSE) +
  scale_color_colorblind()

Mapping vs. setting

What will the following code result in?

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate)
) +
  geom_point(aes(color = homeownership)) +
  geom_smooth(color = "red", se = FALSE) +
  scale_color_colorblind()

Mapping vs. setting

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate)
) +
  geom_point(aes(color = homeownership)) +
  geom_smooth(color = "red", se = FALSE) +
  scale_color_colorblind()

Recap: Aesthetic mappings

• Aesthetic mapping defined at the global level will be used by all geoms for which the aesthetic is defined.

• Aesthetic mapping defined at the local level will be used only by the geoms they’re defined for.

Aside: Legends

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate, color = homeownership, shape = homeownership)
) +
  geom_point() +
  scale_color_colorblind()

Aside: Legends

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate, color = homeownership, shape = homeownership)
) +
  geom_point() +
  scale_color_colorblind() +
  labs(color = "Home ownership")

Aside: Legends

ggplot(
  loan50,
  aes(x = annual_income, y = interest_rate, color = homeownership, shape = homeownership)
) +
  geom_point() +
  scale_color_colorblind() +
  labs(
    color = "Home ownership",
    shape = "Home ownership"
  )

Factors

Factors

• Factors are used for categorical variables – variables that have a fixed and known set of possible values.

• They are also useful when you want to display character vectors in a non-alphabetical order.

R4DS: https://r4ds.hadley.nz/factors.

Bar plot

ggplot(loan50, aes(x = homeownership)) +
  geom_bar()

Bar plot - reordered

loan50 |>
  mutate(homeownership = fct_relevel(homeownership, "mortgage", "rent", "own")) |>
  ggplot(aes(x = homeownership)) +
  geom_bar()

Frequency table

loan50 |>
  count(homeownership)

# A tibble: 3 × 2
  homeownership     n
  <fct>         <int>
1 rent             21
2 mortgage         26
3 own               3

Bar plot - reordered

loan50 |>
  mutate(homeownership = fct_relevel(homeownership, "own", "rent", "mortgage")) |>
  count(homeownership)

# A tibble: 3 × 2
  homeownership     n
  <fct>         <int>
1 own               3
2 rent             21
3 mortgage         26

Under the hood

class(loan50$homeownership)

[1] "factor"

typeof(loan50$homeownership)

[1] "integer"

levels(loan50$homeownership)

[1] "rent"     "mortgage" "own"     

str(loan50$homeownership)

 Factor w/ 3 levels "rent","mortgage",..: 1 1 2 1 2 2 1 2 1 2 ...

Recap: Factors

• The forcats package has a bunch of functions (that start with fct_*()) for dealing with factors and their levels: https://forcats.tidyverse.org/reference/index.html

• Factors and the order of their levels are relevant for displays (tables, plots) and they’ll be relevant for modeling (later in the course)

• factor is a data class

Aside: ==

loan50 |>
  mutate(homeownership_new = if_else(homeownership == "rent", "don't own", homeownership)) |>
  distinct(homeownership, homeownership_new)

# A tibble: 3 × 2
  homeownership homeownership_new
  <fct>         <chr>            
1 rent          don't own        
2 mortgage      mortgage         
3 own           own              

Aside: |

loan50 |>
  mutate(homeownership_new = if_else(homeownership == "rent" | homeownership == "mortgage", "don't own", homeownership)) |>
  distinct(homeownership, homeownership_new)

# A tibble: 3 × 2
  homeownership homeownership_new
  <fct>         <chr>            
1 rent          don't own        
2 mortgage      don't own        
3 own           own              

Aside: |

loan50 |>
  mutate(homeownership_new = if_else(homeownership %in% c("rent", "mortgage"), "don't own", homeownership)) |>
  distinct(homeownership, homeownership_new)

# A tibble: 3 × 2
  homeownership homeownership_new
  <fct>         <chr>            
1 rent          don't own        
2 mortgage      don't own        
3 own           own              

Other questions?