R - Programming with the Tidyverse
1 Installing the ‘tidyverse’ package
Before we have access to the ‘tidyverse’ functions that was mentioned - we will first need to install the package using the install.packages() function. This is a generic function that you can use to install ANY packages you learn about in the future that is part of CRAN. You can install any package by adding the name of the package between quotes, within the parantheses. e.g. ‘tidyverse’
install.packages('tidyverse')
install.packages('janitor')
install.packages('readxl')2 Loading packages
Now that the tidyverse packages have been installed into R, you will need to load the package to get access to the function. You won’t have access to any of the functions until you load the package. There are 2 other packages we will load to help faciliate our analysis.
library(tidyverse)
library(janitor)
library(readxl)3 Case study with the sepsis data
3.1 Reading in the data
## reading in the dataset using the read_excel() function from the readxl package
df <- read_excel('data/sepsis_wide.xlsx', na = 'NA')
## here we are using the janitor package to help us clean the variable names of the data
df <- clean_names(df)3.2 Checking variable data types
Obtaining a glimpse() of the data frame, this allows us to have a quick ‘glimpse’ or look of the data we just imported
glimpse(df)Rows: 455
Columns: 21
$ id <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, …
$ treat <dbl> 0, 1, 0, 1, 0, 1, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 1, 0, …
$ race <dbl> 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 2, 0, …
$ apache <dbl> 27, 14, 33, 3, 5, 13, 34, 11, 25, 20, 21, 14, 19, 23, 22, 16, 1…
$ death <dbl> 1, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 1, …
$ temp0 <dbl> 95.4, 101.6, 101.0, 101.0, 101.4, 100.7, 101.1, 100.9, 100.8, 9…
$ temp1 <dbl> 93.9, 99.0, 98.9, 100.2, 101.4, 101.0, 99.7, 100.3, 99.7, 98.0,…
$ temp2 <dbl> 93.6, 99.8, 97.2, 100.0, 101.8, 101.5, 99.5, 100.7, 99.9, 97.0,…
$ temp3 <dbl> 93.0, 99.8, NA, 99.5, 101.4, 102.1, 99.0, 100.3, 99.7, 99.6, 10…
$ temp4 <dbl> 90.3, 99.6, NA, 99.7, 101.4, 102.3, 99.1, 99.1, 99.0, 99.6, 100…
$ temp5 <dbl> 93.2, 100.2, 94.8, 99.1, 101.0, 99.9, 98.7, 98.9, 98.1, 100.0, …
$ temp6 <dbl> 95.0, 100.2, 95.5, 99.3, 102.6, 100.5, 99.1, 99.3, 98.9, 99.0, …
$ temp7 <dbl> 96.2, 99.0, NA, 99.6, 100.8, 101.5, 98.3, 100.1, 98.1, 100.6, 1…
$ temp8 <dbl> 98.8, 99.9, 97.2, 98.0, 100.6, 101.9, 99.1, 100.3, 98.0, 101.0,…
$ temp9 <dbl> 97.5, 99.8, 99.2, 97.8, 99.8, 99.5, 99.0, 99.3, 98.1, 101.4, 10…
$ temp10 <dbl> 97.8, 99.6, NA, 97.6, 99.6, 100.7, 98.1, 98.7, 98.1, 99.8, 100.…
$ temp11 <dbl> 99.1, 99.4, NA, 98.4, 99.6, 101.3, 99.0, 98.5, 97.8, 103.2, 100…
$ temp12 <dbl> 95.6, 101.2, NA, 98.0, 99.8, 101.5, 98.9, 99.0, 97.8, 97.4, 98.…
$ temp13 <dbl> NA, 101.0, NA, 98.4, 100.8, 101.5, NA, 98.6, 98.0, 102.0, 101.0…
$ temp14 <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ temp15 <dbl> NA, 100.6, NA, 98.8, 100.0, 101.5, NA, 99.0, 98.0, 101.4, 101.0…3.3 Initial pre-processing
We can see that treat, race, and death are incorrectly coded as ‘dbl’, we would like to change them into factors using the factors() function. Since we are replacing the previous values with a new one, we need to use it in conjunction with the mutate function to create a new variable
df <- mutate(df, treat = factor(
treat,
levels = c(0, 1),
labels = c('Control', 'Ibuprofenol')
))
df <- mutate(df, death = factor(
death,
levels = c(0, 1),
labels = c('Alive', 'Died')
))
df <- mutate(df, race = factor(
race,
levels = c(0, 1, 2),
labels = c('White', 'African American', 'Hispanic')
))
## obtain a new glimpse of the data to confirm the variables contain the correct 'typing'
glimpse(df)Rows: 455
Columns: 21
$ id <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, …
$ treat <fct> Control, Ibuprofenol, Control, Ibuprofenol, Control, Ibuprofeno…
$ race <fct> White, African American, African American, White, White, White,…
$ apache <dbl> 27, 14, 33, 3, 5, 13, 34, 11, 25, 20, 21, 14, 19, 23, 22, 16, 1…
$ death <fct> Died, Alive, Died, Alive, Alive, Alive, Died, Alive, Alive, Ali…
$ temp0 <dbl> 95.4, 101.6, 101.0, 101.0, 101.4, 100.7, 101.1, 100.9, 100.8, 9…
$ temp1 <dbl> 93.9, 99.0, 98.9, 100.2, 101.4, 101.0, 99.7, 100.3, 99.7, 98.0,…
$ temp2 <dbl> 93.6, 99.8, 97.2, 100.0, 101.8, 101.5, 99.5, 100.7, 99.9, 97.0,…
$ temp3 <dbl> 93.0, 99.8, NA, 99.5, 101.4, 102.1, 99.0, 100.3, 99.7, 99.6, 10…
$ temp4 <dbl> 90.3, 99.6, NA, 99.7, 101.4, 102.3, 99.1, 99.1, 99.0, 99.6, 100…
$ temp5 <dbl> 93.2, 100.2, 94.8, 99.1, 101.0, 99.9, 98.7, 98.9, 98.1, 100.0, …
$ temp6 <dbl> 95.0, 100.2, 95.5, 99.3, 102.6, 100.5, 99.1, 99.3, 98.9, 99.0, …
$ temp7 <dbl> 96.2, 99.0, NA, 99.6, 100.8, 101.5, 98.3, 100.1, 98.1, 100.6, 1…
$ temp8 <dbl> 98.8, 99.9, 97.2, 98.0, 100.6, 101.9, 99.1, 100.3, 98.0, 101.0,…
$ temp9 <dbl> 97.5, 99.8, 99.2, 97.8, 99.8, 99.5, 99.0, 99.3, 98.1, 101.4, 10…
$ temp10 <dbl> 97.8, 99.6, NA, 97.6, 99.6, 100.7, 98.1, 98.7, 98.1, 99.8, 100.…
$ temp11 <dbl> 99.1, 99.4, NA, 98.4, 99.6, 101.3, 99.0, 98.5, 97.8, 103.2, 100…
$ temp12 <dbl> 95.6, 101.2, NA, 98.0, 99.8, 101.5, 98.9, 99.0, 97.8, 97.4, 98.…
$ temp13 <dbl> NA, 101.0, NA, 98.4, 100.8, 101.5, NA, 98.6, 98.0, 102.0, 101.0…
$ temp14 <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ temp15 <dbl> NA, 100.6, NA, 98.8, 100.0, 101.5, NA, 99.0, 98.0, 101.4, 101.0…3.4 Converting the temperature from F to C
## we can write the formula to convert a single temp from F to C
mutate(df, temp0 = (temp0 - 32) * 5/9)# A tibble: 455 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 93.9 93.6 93 90.3 93.2 95
2 2 Ibuprofen… Afri… 14 Alive 38.7 99 99.8 99.8 99.6 100. 100.
3 3 Control Afri… 33 Died 38.3 98.9 97.2 NA NA 94.8 95.5
4 4 Ibuprofen… White 3 Alive 38.3 100. 100 99.5 99.7 99.1 99.3
5 5 Control White 5 Alive 38.6 101. 102. 101. 101. 101 103.
6 6 Ibuprofen… White 13 Alive 38.2 101 102. 102. 102. 99.9 100.
7 7 Ibuprofen… White 34 Died 38.4 99.7 99.5 99 99.1 98.7 99.1
8 8 Control White 11 Alive 38.3 100. 101. 100. 99.1 98.9 99.3
9 9 Ibuprofen… White 25 Alive 38.2 99.7 99.9 99.7 99 98.1 98.9
10 10 Control White 20 Alive 37.2 98 97 99.6 99.6 100 99
# … with 445 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>## alternatively we can write our own function to do this
fare_to_cel <- function(x) {
(x - 32) * 5/9
}
## and then apply the function to the selected variables using across() within the mutate function use ?across to get more information on how to use this function the below three methods all produce the same results
df %>%
mutate(across(
c(
"temp0",
"temp1",
"temp2",
"temp3",
"temp4",
"temp5",
"temp6",
"temp7",
"temp8",
"temp9",
"temp10",
"temp11",
"temp12",
"temp13",
"temp14",
"temp15"
),
fare_to_cel
))# A tibble: 455 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 34.4 34.2 33.9 32.4 34.0 35
2 2 Ibuprofen… Afri… 14 Alive 38.7 37.2 37.7 37.7 37.6 37.9 37.9
3 3 Control Afri… 33 Died 38.3 37.2 36.2 NA NA 34.9 35.3
4 4 Ibuprofen… White 3 Alive 38.3 37.9 37.8 37.5 37.6 37.3 37.4
5 5 Control White 5 Alive 38.6 38.6 38.8 38.6 38.6 38.3 39.2
6 6 Ibuprofen… White 13 Alive 38.2 38.3 38.6 38.9 39.1 37.7 38.1
7 7 Ibuprofen… White 34 Died 38.4 37.6 37.5 37.2 37.3 37.1 37.3
8 8 Control White 11 Alive 38.3 37.9 38.2 37.9 37.3 37.2 37.4
9 9 Ibuprofen… White 25 Alive 38.2 37.6 37.7 37.6 37.2 36.7 37.2
10 10 Control White 20 Alive 37.2 36.7 36.1 37.6 37.6 37.8 37.2
# … with 445 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>df %>% mutate(across(c(6:21), fare_to_cel))# A tibble: 455 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 34.4 34.2 33.9 32.4 34.0 35
2 2 Ibuprofen… Afri… 14 Alive 38.7 37.2 37.7 37.7 37.6 37.9 37.9
3 3 Control Afri… 33 Died 38.3 37.2 36.2 NA NA 34.9 35.3
4 4 Ibuprofen… White 3 Alive 38.3 37.9 37.8 37.5 37.6 37.3 37.4
5 5 Control White 5 Alive 38.6 38.6 38.8 38.6 38.6 38.3 39.2
6 6 Ibuprofen… White 13 Alive 38.2 38.3 38.6 38.9 39.1 37.7 38.1
7 7 Ibuprofen… White 34 Died 38.4 37.6 37.5 37.2 37.3 37.1 37.3
8 8 Control White 11 Alive 38.3 37.9 38.2 37.9 37.3 37.2 37.4
9 9 Ibuprofen… White 25 Alive 38.2 37.6 37.7 37.6 37.2 36.7 37.2
10 10 Control White 20 Alive 37.2 36.7 36.1 37.6 37.6 37.8 37.2
# … with 445 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>df %>% mutate(across(c(temp0:temp15), fare_to_cel))# A tibble: 455 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 34.4 34.2 33.9 32.4 34.0 35
2 2 Ibuprofen… Afri… 14 Alive 38.7 37.2 37.7 37.7 37.6 37.9 37.9
3 3 Control Afri… 33 Died 38.3 37.2 36.2 NA NA 34.9 35.3
4 4 Ibuprofen… White 3 Alive 38.3 37.9 37.8 37.5 37.6 37.3 37.4
5 5 Control White 5 Alive 38.6 38.6 38.8 38.6 38.6 38.3 39.2
6 6 Ibuprofen… White 13 Alive 38.2 38.3 38.6 38.9 39.1 37.7 38.1
7 7 Ibuprofen… White 34 Died 38.4 37.6 37.5 37.2 37.3 37.1 37.3
8 8 Control White 11 Alive 38.3 37.9 38.2 37.9 37.3 37.2 37.4
9 9 Ibuprofen… White 25 Alive 38.2 37.6 37.7 37.6 37.2 36.7 37.2
10 10 Control White 20 Alive 37.2 36.7 36.1 37.6 37.6 37.8 37.2
# … with 445 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>## if we are content with the changes, we can save and replace our original dataframe with the changes
df <- df %>% mutate(across(c(temp0:temp15), fare_to_cel))3.5 Transforming from wide to long and back
We can use the pivot_longer() and the pivot_wider() function to transpose our data between the two data structure. The pivot_longer() function will transpose a wide data frame to a long data frame. The pivot_wider() function will allow for the opposite, a long dataframe to a wide dataframe.
## here we are creating a new dataframe called 'long'
long <- df %>% pivot_longer(., cols = temp0:temp15, names_to = 'name', values_to = 'value')
long# A tibble: 7,280 × 7
id treat race apache death name value
<dbl> <fct> <fct> <dbl> <fct> <chr> <dbl>
1 1 Control White 27 Died temp0 35.2
2 1 Control White 27 Died temp1 34.4
3 1 Control White 27 Died temp2 34.2
4 1 Control White 27 Died temp3 33.9
5 1 Control White 27 Died temp4 32.4
6 1 Control White 27 Died temp5 34.0
7 1 Control White 27 Died temp6 35
8 1 Control White 27 Died temp7 35.7
9 1 Control White 27 Died temp8 37.1
10 1 Control White 27 Died temp9 36.4
# … with 7,270 more rows## here we are creating a new dataframe called 'wide' by converting the long dataframe back to wide format they are both valid formats we can work with in R
wide <- long %>% pivot_wider(., names_from = 'name', values_from = 'value')
wide# A tibble: 455 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 34.4 34.2 33.9 32.4 34.0 35
2 2 Ibuprofen… Afri… 14 Alive 38.7 37.2 37.7 37.7 37.6 37.9 37.9
3 3 Control Afri… 33 Died 38.3 37.2 36.2 NA NA 34.9 35.3
4 4 Ibuprofen… White 3 Alive 38.3 37.9 37.8 37.5 37.6 37.3 37.4
5 5 Control White 5 Alive 38.6 38.6 38.8 38.6 38.6 38.3 39.2
6 6 Ibuprofen… White 13 Alive 38.2 38.3 38.6 38.9 39.1 37.7 38.1
7 7 Ibuprofen… White 34 Died 38.4 37.6 37.5 37.2 37.3 37.1 37.3
8 8 Control White 11 Alive 38.3 37.9 38.2 37.9 37.3 37.2 37.4
9 9 Ibuprofen… White 25 Alive 38.2 37.6 37.7 37.6 37.2 36.7 37.2
10 10 Control White 20 Alive 37.2 36.7 36.1 37.6 37.6 37.8 37.2
# … with 445 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>3.6 Extracting specific observations
## to extract specific observations we can use the filter() function, and state a criteria. Here we are extracting only patients in the Control group
df %>% filter(., treat == 'Control')# A tibble: 231 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 34.4 34.2 33.9 32.4 34.0 35
2 3 Control African… 33 Died 38.3 37.2 36.2 NA NA 34.9 35.3
3 5 Control White 5 Alive 38.6 38.6 38.8 38.6 38.6 38.3 39.2
4 8 Control White 11 Alive 38.3 37.9 38.2 37.9 37.3 37.2 37.4
5 10 Control White 20 Alive 37.2 36.7 36.1 37.6 37.6 37.8 37.2
6 12 Control White 14 Alive 38.6 38.7 39.2 39.1 39.1 39.7 38.3
7 14 Control White 23 Alive 37.9 37.6 37.4 37.4 37.4 37.6 37.6
8 16 Control White 16 Died 38.1 38.0 37.7 36.9 36.9 36.4 NA
9 17 Control African… 17 Alive 38.1 37.7 38.3 37.8 39.3 37.6 38.2
10 18 Control White 14 Alive 38.7 38.7 38.6 38.3 38.3 38.3 38.4
# … with 221 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>## here we are extracting white control patients
df %>% filter(., treat == 'Control' & race == 'White')# A tibble: 156 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 34.4 34.2 33.9 32.4 34.0 35
2 5 Control White 5 Alive 38.6 38.6 38.8 38.6 38.6 38.3 39.2
3 8 Control White 11 Alive 38.3 37.9 38.2 37.9 37.3 37.2 37.4
4 10 Control White 20 Alive 37.2 36.7 36.1 37.6 37.6 37.8 37.2
5 12 Control White 14 Alive 38.6 38.7 39.2 39.1 39.1 39.7 38.3
6 14 Control White 23 Alive 37.9 37.6 37.4 37.4 37.4 37.6 37.6
7 16 Control White 16 Died 38.1 38.0 37.7 36.9 36.9 36.4 NA
8 18 Control White 14 Alive 38.7 38.7 38.6 38.3 38.3 38.3 38.4
9 21 Control White 24 Died 39.3 39.4 39.8 39.9 38.6 36.8 36.8
10 23 Control White 14 Alive 38.4 38.7 38.2 37.8 38.2 36.8 38.1
# … with 146 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>## here we are extracting white control patients who died
df %>% filter(., treat == 'Control' & race == 'White' & death == 'Died')# A tibble: 54 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 34.4 34.2 33.9 32.4 34.0 35
2 16 Control White 16 Died 38.1 38.0 37.7 36.9 36.9 36.4 NA
3 21 Control White 24 Died 39.3 39.4 39.8 39.9 38.6 36.8 36.8
4 27 Control White 18 Died 38.3 37.6 37.6 36.8 37.7 36.7 38.0
5 56 Control White 18 Died 39.1 39.3 39.5 39.5 39.8 39.6 39.9
6 80 Control White 27 Died 37.7 37.7 37.2 37.9 38.2 37.9 37.8
7 82 Control White 8 Died 38.1 36.4 36.6 34.4 33.0 34.4 34.8
8 89 Control White 21 Died 34.6 37.3 37.4 37.6 37.1 37.4 37.4
9 102 Control White 15 Died 38.6 38.5 38.6 38.3 37.6 37.1 36.7
10 104 Control White 7 Died 38.0 38.0 38.0 38.2 37.7 37.3 39.0
# … with 44 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>## here we are extracting white control patients with apache score greater than 10 or who died
df %>% filter(., (treat == 'Control' & race == 'White' ) & (apache > 10 | death == 'Died'))# A tibble: 121 × 21
id treat race apache death temp0 temp1 temp2 temp3 temp4 temp5 temp6
<dbl> <fct> <fct> <dbl> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 Died 35.2 34.4 34.2 33.9 32.4 34.0 35
2 8 Control White 11 Alive 38.3 37.9 38.2 37.9 37.3 37.2 37.4
3 10 Control White 20 Alive 37.2 36.7 36.1 37.6 37.6 37.8 37.2
4 12 Control White 14 Alive 38.6 38.7 39.2 39.1 39.1 39.7 38.3
5 14 Control White 23 Alive 37.9 37.6 37.4 37.4 37.4 37.6 37.6
6 16 Control White 16 Died 38.1 38.0 37.7 36.9 36.9 36.4 NA
7 18 Control White 14 Alive 38.7 38.7 38.6 38.3 38.3 38.3 38.4
8 21 Control White 24 Died 39.3 39.4 39.8 39.9 38.6 36.8 36.8
9 23 Control White 14 Alive 38.4 38.7 38.2 37.8 38.2 36.8 38.1
10 25 Control White 22 Alive 35.7 36.8 37.4 38.3 38.4 37.9 37.4
# … with 111 more rows, and 9 more variables: temp7 <dbl>, temp8 <dbl>,
# temp9 <dbl>, temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>3.7 Selecting specific columns or variables
## perhaps we wanted to extract a dataframe that only contains id, treat, race, apache, and death
df %>% select(., id, treat, race, apache, death)# A tibble: 455 × 5
id treat race apache death
<dbl> <fct> <fct> <dbl> <fct>
1 1 Control White 27 Died
2 2 Ibuprofenol African American 14 Alive
3 3 Control African American 33 Died
4 4 Ibuprofenol White 3 Alive
5 5 Control White 5 Alive
6 6 Ibuprofenol White 13 Alive
7 7 Ibuprofenol White 34 Died
8 8 Control White 11 Alive
9 9 Ibuprofenol White 25 Alive
10 10 Control White 20 Alive
# … with 445 more rows## perhaps we wanted to extract only the id and the temp variables
df %>% select(., id, temp0:temp15)# A tibble: 455 × 17
id temp0 temp1 temp2 temp3 temp4 temp5 temp6 temp7 temp8 temp9 temp10
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 35.2 34.4 34.2 33.9 32.4 34.0 35 35.7 37.1 36.4 36.6
2 2 38.7 37.2 37.7 37.7 37.6 37.9 37.9 37.2 37.7 37.7 37.6
3 3 38.3 37.2 36.2 NA NA 34.9 35.3 NA 36.2 37.3 NA
4 4 38.3 37.9 37.8 37.5 37.6 37.3 37.4 37.6 36.7 36.6 36.4
5 5 38.6 38.6 38.8 38.6 38.6 38.3 39.2 38.2 38.1 37.7 37.6
6 6 38.2 38.3 38.6 38.9 39.1 37.7 38.1 38.6 38.8 37.5 38.2
7 7 38.4 37.6 37.5 37.2 37.3 37.1 37.3 36.8 37.3 37.2 36.7
8 8 38.3 37.9 38.2 37.9 37.3 37.2 37.4 37.8 37.9 37.4 37.1
9 9 38.2 37.6 37.7 37.6 37.2 36.7 37.2 36.7 36.7 36.7 36.7
10 10 37.2 36.7 36.1 37.6 37.6 37.8 37.2 38.1 38.3 38.6 37.7
# … with 445 more rows, and 5 more variables: temp11 <dbl>, temp12 <dbl>,
# temp13 <dbl>, temp14 <dbl>, temp15 <dbl>## perhaps we wanted to extract all the variables except outcome e.g. death
df %>% select(., -death)# A tibble: 455 × 20
id treat race apache temp0 temp1 temp2 temp3 temp4 temp5 temp6 temp7
<dbl> <fct> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 1 Control White 27 35.2 34.4 34.2 33.9 32.4 34.0 35 35.7
2 2 Ibuprofen… Afri… 14 38.7 37.2 37.7 37.7 37.6 37.9 37.9 37.2
3 3 Control Afri… 33 38.3 37.2 36.2 NA NA 34.9 35.3 NA
4 4 Ibuprofen… White 3 38.3 37.9 37.8 37.5 37.6 37.3 37.4 37.6
5 5 Control White 5 38.6 38.6 38.8 38.6 38.6 38.3 39.2 38.2
6 6 Ibuprofen… White 13 38.2 38.3 38.6 38.9 39.1 37.7 38.1 38.6
7 7 Ibuprofen… White 34 38.4 37.6 37.5 37.2 37.3 37.1 37.3 36.8
8 8 Control White 11 38.3 37.9 38.2 37.9 37.3 37.2 37.4 37.8
9 9 Ibuprofen… White 25 38.2 37.6 37.7 37.6 37.2 36.7 37.2 36.7
10 10 Control White 20 37.2 36.7 36.1 37.6 37.6 37.8 37.2 38.1
# … with 445 more rows, and 8 more variables: temp8 <dbl>, temp9 <dbl>,
# temp10 <dbl>, temp11 <dbl>, temp12 <dbl>, temp13 <dbl>, temp14 <dbl>,
# temp15 <dbl>## we can also use a numerical indicator of the column number to select a range of columns, e.g the first 5 columns
df %>% select(., 1:5)# A tibble: 455 × 5
id treat race apache death
<dbl> <fct> <fct> <dbl> <fct>
1 1 Control White 27 Died
2 2 Ibuprofenol African American 14 Alive
3 3 Control African American 33 Died
4 4 Ibuprofenol White 3 Alive
5 5 Control White 5 Alive
6 6 Ibuprofenol White 13 Alive
7 7 Ibuprofenol White 34 Died
8 8 Control White 11 Alive
9 9 Ibuprofenol White 25 Alive
10 10 Control White 20 Alive
# … with 445 more rows3.7.1 Selecting using helper functions
## let's select variables using starts_with() helper function
df %>% select(., starts_with('ra'))# A tibble: 455 × 1
race
<fct>
1 White
2 African American
3 African American
4 White
5 White
6 White
7 White
8 White
9 White
10 White
# … with 445 more rows## let's select variables using ends_with() helper function
df %>% select(., ends_with('e'))# A tibble: 455 × 2
race apache
<fct> <dbl>
1 White 27
2 African American 14
3 African American 33
4 White 3
5 White 5
6 White 13
7 White 34
8 White 11
9 White 25
10 White 20
# … with 445 more rows## let's select variables using contains() helper function
df %>% select(., contains('temp'))# A tibble: 455 × 16
temp0 temp1 temp2 temp3 temp4 temp5 temp6 temp7 temp8 temp9 temp10 temp11
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 35.2 34.4 34.2 33.9 32.4 34.0 35 35.7 37.1 36.4 36.6 37.3
2 38.7 37.2 37.7 37.7 37.6 37.9 37.9 37.2 37.7 37.7 37.6 37.4
3 38.3 37.2 36.2 NA NA 34.9 35.3 NA 36.2 37.3 NA NA
4 38.3 37.9 37.8 37.5 37.6 37.3 37.4 37.6 36.7 36.6 36.4 36.9
5 38.6 38.6 38.8 38.6 38.6 38.3 39.2 38.2 38.1 37.7 37.6 37.6
6 38.2 38.3 38.6 38.9 39.1 37.7 38.1 38.6 38.8 37.5 38.2 38.5
7 38.4 37.6 37.5 37.2 37.3 37.1 37.3 36.8 37.3 37.2 36.7 37.2
8 38.3 37.9 38.2 37.9 37.3 37.2 37.4 37.8 37.9 37.4 37.1 36.9
9 38.2 37.6 37.7 37.6 37.2 36.7 37.2 36.7 36.7 36.7 36.7 36.6
10 37.2 36.7 36.1 37.6 37.6 37.8 37.2 38.1 38.3 38.6 37.7 39.6
# … with 445 more rows, and 4 more variables: temp12 <dbl>, temp13 <dbl>,
# temp14 <dbl>, temp15 <dbl>## let's select variables using num_range() helper function
df %>% select(., num_range('temp', 0:5))# A tibble: 455 × 6
temp0 temp1 temp2 temp3 temp4 temp5
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 35.2 34.4 34.2 33.9 32.4 34.0
2 38.7 37.2 37.7 37.7 37.6 37.9
3 38.3 37.2 36.2 NA NA 34.9
4 38.3 37.9 37.8 37.5 37.6 37.3
5 38.6 38.6 38.8 38.6 38.6 38.3
6 38.2 38.3 38.6 38.9 39.1 37.7
7 38.4 37.6 37.5 37.2 37.3 37.1
8 38.3 37.9 38.2 37.9 37.3 37.2
9 38.2 37.6 37.7 37.6 37.2 36.7
10 37.2 36.7 36.1 37.6 37.6 37.8
# … with 445 more rows3.8 Summarizing data with dplyr
## lets taking a simple count of the dataframe
df %>% summarise(., n = n())# A tibble: 1 × 1
n
<int>
1 455## the summzarise function becomes very powerful when you use it in conjuntion with the group_by() function lets take the count of control and ibuprofenol patients in our dataset
df %>%
group_by(treat) %>%
summarise(., n = n(), .groups = 'drop')# A tibble: 2 × 2
treat n
<fct> <int>
1 Control 231
2 Ibuprofenol 224## we can expand this further by including more factors in group_by() here lets take the count of people who died among the different arms
df %>%
group_by(., treat, death) %>%
summarise(., n = n(), .groups = 'drop')# A tibble: 4 × 3
treat death n
<fct> <fct> <int>
1 Control Alive 139
2 Control Died 92
3 Ibuprofenol Alive 140
4 Ibuprofenol Died 84## we can also further expand this by adding more functions into the summarise function
df %>%
group_by(., treat, death) %>%
summarise(
.,
mean_temp0 = mean(temp0, na.rm = T),
sd_temp0 = sd(temp0, na.rm = T),
.groups = 'drop'
)# A tibble: 4 × 4
treat death mean_temp0 sd_temp0
<fct> <fct> <dbl> <dbl>
1 Control Alive 38.2 0.765
2 Control Died 37.8 1.35
3 Ibuprofenol Alive 38.1 1.09
4 Ibuprofenol Died 37.8 1.34 ## here we are using a across which allows us to easily summarise many columns at once, with a list of functions
df %>%
group_by(., treat, death) %>%
summarise(across(
c(temp0:temp15),
list('mean' = mean, 'sd' = sd),
na.rm = T,
.names = '{.fn}_{.col}'
),
.groups = 'drop')# A tibble: 4 × 34
treat death mean_…¹ sd_te…² mean_…³ sd_te…⁴ mean_…⁵ sd_te…⁶ mean_…⁷ sd_te…⁸
<fct> <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 Control Alive 38.2 0.765 38.1 0.923 38.0 0.764 38.1 0.799
2 Control Died 37.8 1.35 37.5 1.37 37.5 1.29 37.5 1.32
3 Ibuprof… Alive 38.1 1.09 37.6 0.906 37.2 0.883 37.1 0.931
4 Ibuprof… Died 37.8 1.34 37.4 1.22 37.1 1.22 36.8 1.04
# … with 24 more variables: mean_temp4 <dbl>, sd_temp4 <dbl>, mean_temp5 <dbl>,
# sd_temp5 <dbl>, mean_temp6 <dbl>, sd_temp6 <dbl>, mean_temp7 <dbl>,
# sd_temp7 <dbl>, mean_temp8 <dbl>, sd_temp8 <dbl>, mean_temp9 <dbl>,
# sd_temp9 <dbl>, mean_temp10 <dbl>, sd_temp10 <dbl>, mean_temp11 <dbl>,
# sd_temp11 <dbl>, mean_temp12 <dbl>, sd_temp12 <dbl>, mean_temp13 <dbl>,
# sd_temp13 <dbl>, mean_temp14 <dbl>, sd_temp14 <dbl>, mean_temp15 <dbl>,
# sd_temp15 <dbl>, and abbreviated variable names ¹mean_temp0, ²sd_temp0, …