R pour l’analyse de données
Pierre Gloaguen
November, 2020
1 Prérequis supposés
1.1 Prerequis
- Savoir lire l’anglais
- Bases de
R- Créer des objets
- Opérations de base
- Objets de base en
R- Vecteurs
- Listes
- Data.frame
- Fonctions de bases
- Import de données
1.2 Références utiles
R for Data Science (G. Grolemund and H. Wickham)
R Markdown: The definitive guide (Y. Xie, J.J. Allaire and G. Grolemund)
ggplot2: Elegant Graphics for Data Analysis (H. Wickham)
Les deux premiers sont disponible en ligne
2 Tidyverse
2.1 What is tidyverse?
Set of packages for tidy and unified data processing, visualization and modelling.
Consists in a set of method having unified structure.
Involves slight modifications in classical
Rgrammar.This grammar becomes dominant, and highly documented.
2.2 Loading tidyverse
library(tidyverse)- Load a lot of packages
- Some of them are often updated! Watch for updates!
3 The tibble format
3.1 tibble, a nicer data.frame
- The
tibbleis a “modern” version of thedata.frame
iris_tibble <- as_tibble(iris) # Natural conversion- The printing is naturally cropped, giving types, number of rows, cols
iris_tibble# A tibble: 150 x 5
Sepal.Length Sepal.Width Petal.Length Petal.Width Species
<dbl> <dbl> <dbl> <dbl> <fct>
1 5.1 3.5 1.4 0.2 setosa
2 4.9 3 1.4 0.2 setosa
3 4.7 3.2 1.3 0.2 setosa
4 4.6 3.1 1.5 0.2 setosa
5 5 3.6 1.4 0.2 setosa
6 5.4 3.9 1.7 0.4 setosa
7 4.6 3.4 1.4 0.3 setosa
8 5 3.4 1.5 0.2 setosa
9 4.4 2.9 1.4 0.2 setosa
10 4.9 3.1 1.5 0.1 setosa
# … with 140 more rows3.2 Creating a tibble
The function tibble creates naturally a tibble.
my_tibble <- tibble(Nom = c("Alice", "Bob", "Claire"),
Age = c(10, 25, 30))It can be neatly printed with knitr::kable, like a data.frame
my_tibble# A tibble: 3 x 2
Nom Age
<chr> <dbl>
1 Alice 10
2 Bob 25
3 Claire 303.3 Extracting in tibble
Exactly as in data.frame
my_tibble[1] # Returns a tibble
my_tibble[, "Nom"]# Returns a tibble. Equivalent to my_tibble["Nom"]
# For a data.frame, this last line would return a vector, not a data.frameTo extract columns, the pull function is introduced (also works with data.frame).
dplyr::pull(my_tibble, Nom)# The pull function extracts a column to a vector[1] "Alice" "Bob" "Claire"my_tibble$Nom # But the old fashion way still works[1] "Alice" "Bob" "Claire"3.4 Row names in tibble
For programming reasons, tibble can’t have row names
head(swiss[, 1:3], n = 2) Fertility Agriculture Examination
Courtelary 80.2 17.0 15
Delemont 83.1 45.1 6print(as_tibble(swiss[, 1:3]), n = 2)# A tibble: 47 x 3
Fertility Agriculture Examination
<dbl> <dbl> <int>
1 80.2 17 15
2 83.1 45.1 6
# … with 45 more rowsprint(as_tibble(swiss[, 1:3], rownames = "Province"), n = 2) # Creating a new column# A tibble: 47 x 4
Province Fertility Agriculture Examination
<chr> <dbl> <dbl> <int>
1 Courtelary 80.2 17 15
2 Delemont 83.1 45.1 6
# … with 45 more rows4 %>% For sequential data processing
4.1 diamonds data set
First, we load a toy data set
data(diamonds) # Load the diamond data set# A tibble: 53,940 x 10
carat cut color clarity depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43
2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
4 0.29 Premium I VS2 62.4 58 334 4.2 4.23 2.63
5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75
6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48
7 0.24 Very Good I VVS1 62.3 57 336 3.95 3.98 2.47
8 0.26 Very Good H SI1 61.9 55 337 4.07 4.11 2.53
9 0.22 Fair E VS2 65.1 61 337 3.87 3.78 2.49
10 0.23 Very Good H VS1 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows4.2 Example of sequential processing
- Imagine we want to compute the mean (omitting
NAs) price of diamonds having aGoodcut and a color of typeD.
There is the one-line way, rather cumbersome
mean(diamonds[diamonds$cut == "Good" & diamonds$color == "D", ]$price, na.rm = T)[1] 3405.382There is the way in multiple lines, creating intermediary objects
condition <- diamonds$cut == "Good" & diamonds$color == "D"
sub_prices <- diamonds[condition, ]$price # could have use pull
mean(sub_prices, na.rm = T)4.3 Tidier sequential processing using %>%
The “pipe” instruction %>% allows to write this sequential instruction in an easy to read way, without creating intermediary objects.
x %>% f()is equivalent tof(x)x %>% f(y)is equivalent tof(x, y)When you read code,
%>%is pronounced “then”
# Same example as before
diamonds %>% # We take the data, then
dplyr::filter(cut == "Good", color == "D") %>% # Subsetting (using dplyr)
dplyr::pull(price) %>% # extract the price
mean(na.rm = T) # compute the mean, omitting the NAs[1] 3405.382- The result of the previous treatment is set as the first default argument of the next function
- No redundancy of the diamonds using
$ - The keyboard shortcut in
Rstudiofor%>%isCtrl + Maj + M - Note that the indenting naturally allows commenting
4.4 More about the %>%
If you need to specify specifically which argument you want your input to go to , you can use the dot “.”
# Adjusting a linear regression on diamonds price w.r.t. carat
diamonds %>%
lm(data = ., formula = price ~ carat) # The "." refers to entering argumentSome advices (from the bible of tidyverse)
Use
%>%to emphasize a sequence of actions, rather than the object that the actions are being performed on.Avoid using the pipe when:
- You need to manipulate more than one object at a time. Reserve pipes for a sequence of steps applied to one primary object.
- There are meaningful intermediate objects that could be given informative names.
5 Manipulating data
5.1 The dplyr package
library(dplyr)dplyr is a package (part of tidyverse) which allows you to solve the vast majority of your data-manipulation challenge:
create variables
pick variables
reorder observations
pick observations
create summaries
…
Functions in this package are verbs and have consistent structures.
5.2 Extracting lines on number: slice()
- You can extract lines (or slicing the data) using
slice() - The results remains a
tibble(or adata.frame)
diamonds %>%
slice(c(3, 8)) # Select lines 3 and 8# A tibble: 2 x 10
carat cut color clarity depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
2 0.26 Very Good H SI1 61.9 55 337 4.07 4.11 2.535.3 Conditional extraction: filter()
dplyr::filterallows to return rows with matching conditions
data(diamonds, package = "ggplot2") # data set use
diamonds %>%
dplyr::filter(color == "D", clarity == "SI1")# the , is equivalent to an "&"# A tibble: 2,083 x 10
carat cut color clarity depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.3 Premium D SI1 62.6 59 552 4.23 4.27 2.66
2 0.3 Ideal D SI1 62.5 57 552 4.29 4.32 2.69
3 0.3 Ideal D SI1 62.1 56 552 4.3 4.33 2.68
4 0.75 Very Good D SI1 63.2 56 2760 5.8 5.75 3.65
5 0.71 Very Good D SI1 63.6 58 2764 5.64 5.68 3.6
6 0.71 Ideal D SI1 61.9 59 2764 5.69 5.72 3.53
7 0.73 Very Good D SI1 60.2 56 2768 5.83 5.87 3.52
8 0.7 Very Good D SI1 61.1 58 2768 5.66 5.73 3.48
9 0.72 Ideal D SI1 60.8 57 2782 5.76 5.75 3.5
10 0.72 Premium D SI1 62.7 59 2782 5.73 5.69 3.58
# … with 2,073 more rows# Equivalent to the one before
diamonds %>%
dplyr::filter(color == "D" & clarity == "SI1")# A tibble: 2,083 x 10
carat cut color clarity depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.3 Premium D SI1 62.6 59 552 4.23 4.27 2.66
2 0.3 Ideal D SI1 62.5 57 552 4.29 4.32 2.69
3 0.3 Ideal D SI1 62.1 56 552 4.3 4.33 2.68
4 0.75 Very Good D SI1 63.2 56 2760 5.8 5.75 3.65
5 0.71 Very Good D SI1 63.6 58 2764 5.64 5.68 3.6
6 0.71 Ideal D SI1 61.9 59 2764 5.69 5.72 3.53
7 0.73 Very Good D SI1 60.2 56 2768 5.83 5.87 3.52
8 0.7 Very Good D SI1 61.1 58 2768 5.66 5.73 3.48
9 0.72 Ideal D SI1 60.8 57 2782 5.76 5.75 3.5
10 0.72 Premium D SI1 62.7 59 2782 5.73 5.69 3.58
# … with 2,073 more rows# One with a "or" (using the in)
diamonds %>%
dplyr::filter(color %in% c("D", "E"), clarity == "SI1")# A tibble: 4,509 x 10
carat cut color clarity depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
2 0.3 Premium D SI1 62.6 59 552 4.23 4.27 2.66
3 0.3 Ideal D SI1 62.5 57 552 4.29 4.32 2.69
4 0.3 Ideal D SI1 62.1 56 552 4.3 4.33 2.68
5 0.7 Ideal E SI1 62.5 57 2757 5.7 5.72 3.57
6 0.73 Very Good E SI1 61.6 59 2760 5.77 5.78 3.56
7 0.75 Very Good D SI1 63.2 56 2760 5.8 5.75 3.65
8 0.75 Premium E SI1 59.9 54 2760 6 5.96 3.58
9 0.74 Ideal E SI1 62.3 54 2762 5.8 5.83 3.62
10 0.71 Very Good D SI1 63.6 58 2764 5.64 5.68 3.6
# … with 4,499 more rows5.4 Choosing among columns select()
5.4.1 Selecting on column number
- The basic
selectworks on column number.
diamonds %>%
select(1, 4) # Selections des colonnes 1 et 4# A tibble: 53,940 x 2
carat clarity
<dbl> <ord>
1 0.23 SI2
2 0.21 SI1
3 0.23 VS1
4 0.29 VS2
5 0.31 SI2
6 0.24 VVS2
7 0.24 VVS1
8 0.26 SI1
9 0.22 VS2
10 0.23 VS1
# … with 53,930 more rows5.4.2 Selecting on names
- The basic
selectworks on column names.
diamonds %>%
select(carat, cut, depth)# A tibble: 53,940 x 3
carat cut depth
<dbl> <ord> <dbl>
1 0.23 Ideal 61.5
2 0.21 Premium 59.8
3 0.23 Good 56.9
4 0.29 Premium 62.4
5 0.31 Good 63.3
6 0.24 Very Good 62.8
7 0.24 Very Good 62.3
8 0.26 Very Good 61.9
9 0.22 Fair 65.1
10 0.23 Very Good 59.4
# … with 53,930 more rows5.4.3 Removing column
One can remove column using select, by adding - in front of the column name:
diamonds %>%
select(-carat, -cut) # Select all columns but carat and cut# A tibble: 53,940 x 8
color clarity depth table price x y z
<ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 E SI2 61.5 55 326 3.95 3.98 2.43
2 E SI1 59.8 61 326 3.89 3.84 2.31
3 E VS1 56.9 65 327 4.05 4.07 2.31
4 I VS2 62.4 58 334 4.2 4.23 2.63
5 J SI2 63.3 58 335 4.34 4.35 2.75
6 J VVS2 62.8 57 336 3.94 3.96 2.48
7 I VVS1 62.3 57 336 3.95 3.98 2.47
8 H SI1 61.9 55 337 4.07 4.11 2.53
9 E VS2 65.1 61 337 3.87 3.78 2.49
10 H VS1 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows5.4.4 Selecting on a condition select_if()
- One can select on column condition with
select_if - Put in argument a function on column that returns a TRUE/FALSE
diamonds %>%
select_if(.predicate = is.factor) # Select only factor columns, the .predicate is optional# A tibble: 53,940 x 3
cut color clarity
<ord> <ord> <ord>
1 Ideal E SI2
2 Premium E SI1
3 Good E VS1
4 Premium I VS2
5 Good J SI2
6 Very Good J VVS2
7 Very Good I VVS1
8 Very Good H SI1
9 Fair E VS2
10 Very Good H VS1
# … with 53,930 more rowsdiamonds %>%
select_if(is.numeric) # Select only numeric columns# A tibble: 53,940 x 7
carat depth table price x y z
<dbl> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.23 61.5 55 326 3.95 3.98 2.43
2 0.21 59.8 61 326 3.89 3.84 2.31
3 0.23 56.9 65 327 4.05 4.07 2.31
4 0.29 62.4 58 334 4.2 4.23 2.63
5 0.31 63.3 58 335 4.34 4.35 2.75
6 0.24 62.8 57 336 3.94 3.96 2.48
7 0.24 62.3 57 336 3.95 3.98 2.47
8 0.26 61.9 55 337 4.07 4.11 2.53
9 0.22 65.1 61 337 3.87 3.78 2.49
10 0.23 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows5.4.5 Selecting on a characteristic select_at()
- One can select on position using
select_at. It uses thedplyrfunctionvarsthat allows to enclose text expressions.
diamonds %>%
select_at(.vars = vars(starts_with("c"))) # Select columns starting with "c"# A tibble: 53,940 x 4
carat cut color clarity
<dbl> <ord> <ord> <ord>
1 0.23 Ideal E SI2
2 0.21 Premium E SI1
3 0.23 Good E VS1
4 0.29 Premium I VS2
5 0.31 Good J SI2
6 0.24 Very Good J VVS2
7 0.24 Very Good I VVS1
8 0.26 Very Good H SI1
9 0.22 Fair E VS2
10 0.23 Very Good H VS1
# … with 53,930 more rows5.5 Extracting columns with pull()
- Function to extract a column (by its name or number) as a vector
diamonds %>%
pull(cut) %>% # extract the cut variable by name as a vector
head(n = 5) # only print first 5 values[1] Ideal Premium Good Premium Good
Levels: Fair < Good < Very Good < Premium < Idealdiamonds %>%
pull(2) %>% # extract the cut by number variable as a vector
head(n = 5)[1] Ideal Premium Good Premium Good
Levels: Fair < Good < Very Good < Premium < Ideal5.6 Renaming columns rename()
5.6.1 Basic renaming
- You might want to rename column(s)
diamonds %>%
rename(length = x, # Always New = old
width = y) # You can rename multiple columns at once# A tibble: 53,940 x 10
carat cut color clarity depth table price length width z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43
2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
4 0.29 Premium I VS2 62.4 58 334 4.2 4.23 2.63
5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75
6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48
7 0.24 Very Good I VVS1 62.3 57 336 3.95 3.98 2.47
8 0.26 Very Good H SI1 61.9 55 337 4.07 4.11 2.53
9 0.22 Fair E VS2 65.1 61 337 3.87 3.78 2.49
10 0.23 Very Good H VS1 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows5.6.2 Renaming on condition rename_if()
- In the same way, a
rename_ifbase on a condition - Require as argument a function to apply to each selected column.
# Add the "_fact" suffix to every factor column name
diamonds %>%
rename_if(.predicate = is.factor, .funs = function(x) paste0(x, "_fact"))# A tibble: 53,940 x 10
carat cut_fact color_fact clarity_fact depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43
2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
4 0.29 Premium I VS2 62.4 58 334 4.2 4.23 2.63
5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75
6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48
7 0.24 Very Good I VVS1 62.3 57 336 3.95 3.98 2.47
8 0.26 Very Good H SI1 61.9 55 337 4.07 4.11 2.53
9 0.22 Fair E VS2 65.1 61 337 3.87 3.78 2.49
10 0.23 Very Good H VS1 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows5.6.3 Renaming on a characteristic rename_at()
- As well as a
rename_at() - Require as argument a function to apply to each selected column.
# Putting in upper case columns starting with "c"
diamonds %>%
rename_at(.vars = vars(starts_with("c")), .funs = toupper)# A tibble: 53,940 x 10
CARAT CUT COLOR CLARITY depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43
2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
4 0.29 Premium I VS2 62.4 58 334 4.2 4.23 2.63
5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75
6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48
7 0.24 Very Good I VVS1 62.3 57 336 3.95 3.98 2.47
8 0.26 Very Good H SI1 61.9 55 337 4.07 4.11 2.53
9 0.22 Fair E VS2 65.1 61 337 3.87 3.78 2.49
10 0.23 Very Good H VS1 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows5.7 Create or modify a column: mutate()
5.7.1 Basic modication
- You can modify a column
diamonds %>%
mutate(cut = factor(cut, labels = LETTERS[5:1])) # modifying labels of cut# A tibble: 53,940 x 10
carat cut color clarity depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.23 A E SI2 61.5 55 326 3.95 3.98 2.43
2 0.21 B E SI1 59.8 61 326 3.89 3.84 2.31
3 0.23 D E VS1 56.9 65 327 4.05 4.07 2.31
4 0.29 B I VS2 62.4 58 334 4.2 4.23 2.63
5 0.31 D J SI2 63.3 58 335 4.34 4.35 2.75
6 0.24 C J VVS2 62.8 57 336 3.94 3.96 2.48
7 0.24 C I VVS1 62.3 57 336 3.95 3.98 2.47
8 0.26 C H SI1 61.9 55 337 4.07 4.11 2.53
9 0.22 E E VS2 65.1 61 337 3.87 3.78 2.49
10 0.23 C H VS1 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows- The same function also creates columns
diamonds %>%
mutate(z_square = z^2) # creates a new column as function of an existing one# A tibble: 53,940 x 11
carat cut color clarity depth table price x y z z_square
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43 5.90
2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31 5.34
3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31 5.34
4 0.29 Premium I VS2 62.4 58 334 4.2 4.23 2.63 6.92
5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75 7.56
6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48 6.15
7 0.24 Very Good I VVS1 62.3 57 336 3.95 3.98 2.47 6.10
8 0.26 Very Good H SI1 61.9 55 337 4.07 4.11 2.53 6.40
9 0.22 Fair E VS2 65.1 61 337 3.87 3.78 2.49 6.20
10 0.23 Very Good H VS1 59.4 61 338 4 4.05 2.39 5.71
# … with 53,930 more rows- You can create/modify multiple columns at once
- This will be done sequentially!
diamonds %>%
mutate(cut = factor(cut, labels = LETTERS[5:1]), # modifying labels of cut
y = 2 * y, # doubling the width
z_square = z^2, # creating a column z_square
cut = z_square) # This modification overrides the first one# A tibble: 53,940 x 11
carat cut color clarity depth table price x y z z_square
<dbl> <dbl> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl> <dbl>
1 0.23 5.90 E SI2 61.5 55 326 3.95 7.96 2.43 5.90
2 0.21 5.34 E SI1 59.8 61 326 3.89 7.68 2.31 5.34
3 0.23 5.34 E VS1 56.9 65 327 4.05 8.14 2.31 5.34
4 0.29 6.92 I VS2 62.4 58 334 4.2 8.46 2.63 6.92
5 0.31 7.56 J SI2 63.3 58 335 4.34 8.7 2.75 7.56
6 0.24 6.15 J VVS2 62.8 57 336 3.94 7.92 2.48 6.15
7 0.24 6.10 I VVS1 62.3 57 336 3.95 7.96 2.47 6.10
8 0.26 6.40 H SI1 61.9 55 337 4.07 8.22 2.53 6.40
9 0.22 6.20 E VS2 65.1 61 337 3.87 7.56 2.49 6.20
10 0.23 5.71 H VS1 59.4 61 338 4 8.1 2.39 5.71
# … with 53,930 more rows5.7.2 Modifying columns on a condition mutate_if()
- Again, you can change on a condition
mutate_if
diamonds %>%
mutate_if(is.factor, as.character) # Transform factor columns to character# A tibble: 53,940 x 10
carat cut color clarity depth table price x y z
<dbl> <chr> <chr> <chr> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43
2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
4 0.29 Premium I VS2 62.4 58 334 4.2 4.23 2.63
5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75
6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48
7 0.24 Very Good I VVS1 62.3 57 336 3.95 3.98 2.47
8 0.26 Very Good H SI1 61.9 55 337 4.07 4.11 2.53
9 0.22 Fair E VS2 65.1 61 337 3.87 3.78 2.49
10 0.23 Very Good H VS1 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows- Check the type of
cutcolorandclarity!
5.7.3 Modifying columns on a characteristic mutate_at()
- And on position,
mutate_at
diamonds %>%
mutate_at(vars(starts_with("c")),
.funs = function(x) 0) # Put columns to 0# A tibble: 53,940 x 10
carat cut color clarity depth table price x y z
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0 0 0 0 61.5 55 326 3.95 3.98 2.43
2 0 0 0 0 59.8 61 326 3.89 3.84 2.31
3 0 0 0 0 56.9 65 327 4.05 4.07 2.31
4 0 0 0 0 62.4 58 334 4.2 4.23 2.63
5 0 0 0 0 63.3 58 335 4.34 4.35 2.75
6 0 0 0 0 62.8 57 336 3.94 3.96 2.48
7 0 0 0 0 62.3 57 336 3.95 3.98 2.47
8 0 0 0 0 61.9 55 337 4.07 4.11 2.53
9 0 0 0 0 65.1 61 337 3.87 3.78 2.49
10 0 0 0 0 59.4 61 338 4 4.05 2.39
# … with 53,930 more rows5.8 Sorting data: arrange()
diamonds %>%
arrange(carat, depth) # Sorting on carat, then on depth# A tibble: 53,940 x 10
carat cut color clarity depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 0.2 Premium E VS2 59 60 367 3.81 3.78 2.24
2 0.2 Premium E VS2 59.7 62 367 3.84 3.8 2.28
3 0.2 Ideal E VS2 59.7 55 367 3.86 3.84 2.3
4 0.2 Premium E VS2 59.8 62 367 3.79 3.77 2.26
5 0.2 Premium E SI2 60.2 62 345 3.79 3.75 2.27
6 0.2 Premium E VS2 61.1 59 367 3.81 3.78 2.32
7 0.2 Ideal D VS2 61.5 57 367 3.81 3.77 2.33
8 0.2 Premium D VS2 61.7 60 367 3.77 3.72 2.31
9 0.2 Ideal E VS2 62.2 57 367 3.76 3.73 2.33
10 0.2 Premium D VS2 62.3 60 367 3.73 3.68 2.31
# … with 53,930 more rows# Decreasing sort uses the function desc()
diamonds %>%
arrange(desc(carat, depth)) # Decreasing on carat, increasing on depth# A tibble: 53,940 x 10
carat cut color clarity depth table price x y z
<dbl> <ord> <ord> <ord> <dbl> <dbl> <int> <dbl> <dbl> <dbl>
1 5.01 Fair J I1 65.5 59 18018 10.7 10.5 6.98
2 4.5 Fair J I1 65.8 58 18531 10.2 10.2 6.72
3 4.13 Fair H I1 64.8 61 17329 10 9.85 6.43
4 4.01 Premium I I1 61 61 15223 10.1 10.1 6.17
5 4.01 Premium J I1 62.5 62 15223 10.0 9.94 6.24
6 4 Very Good I I1 63.3 58 15984 10.0 9.94 6.31
7 3.67 Premium I I1 62.4 56 16193 9.86 9.81 6.13
8 3.65 Fair H I1 67.1 53 11668 9.53 9.48 6.38
9 3.51 Premium J VS2 62.5 59 18701 9.66 9.63 6.03
10 3.5 Ideal H I1 62.8 57 12587 9.65 9.59 6.03
# … with 53,930 more rows5.9 Summarising data summarise()
- Useful to summarize a column to a single number.
- Returns a
tibble.
diamonds %>%
summarise(mean_carat = mean(carat),
var_carat = var(carat),
number_diamonds = n(), # dplyr::n() equivalent to nrow(.)
number_distinct_carats = n_distinct(carat))# ddplyr::n_distinct()# A tibble: 1 x 4
mean_carat var_carat number_diamonds number_distinct_carats
<dbl> <dbl> <int> <int>
1 0.798 0.225 53940 2735.9.1 Summarising columns on condition summarise_if()
- Of course,
summarise_if()andsummarise_at()exists!
# Compute the mean of all numeric variables
diamonds %>%
summarise_if(is.numeric, mean)# A tibble: 1 x 7
carat depth table price x y z
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 0.798 61.7 57.5 3933. 5.73 5.73 3.54diamonds %>%
summarise_at(.vars = vars(starts_with("c")), # For columns starting by c
# count number of unique elements in the column
.funs = list(n_unique = function(x) length(unique(x))))# A tibble: 1 x 4
carat_n_unique cut_n_unique color_n_unique clarity_n_unique
<int> <int> <int> <int>
1 273 5 7 85.10 Grouping data group_by()
- Often, you might compute statistics depending on groups.
group_bywiill allow to perform the summarise per factor of a group
diamonds %>%
group_by(cut) %>% # depending on the cut
summarise(Count = n(),# Number of diamonds per group
Mean_price = mean(price),# Mean price per group
Mean_carat = var(carat))# A tibble: 5 x 4
cut Count Mean_price Mean_carat
<ord> <int> <dbl> <dbl>
1 Fair 1610 4359. 0.267
2 Good 4906 3929. 0.206
3 Very Good 12082 3982. 0.211
4 Premium 13791 4584. 0.265
5 Ideal 21551 3458. 0.187- In practice
group_bychanges the waydplyrfunctions operate on the data. They will operate groupwise.
5.11 Exercise
In the diamonds data set
- Rename the
x,y,zvariables aslength,width,height. - Put their values in cm instead of mm.
- Create an object
diamonds_modifcontaining these three columns and theclarity,cut,caratandprice, when for the cut isPremiumorIdeal. - For each combination
cut/color, compute the mean of the three variableslength,width,height. Keep this summary in an objectdim_summary.
5.12 Exercise (Solution)
- First we create a
diamonds_modifas wanted.
diamonds_modif <- diamonds %>%
rename(length = x, width = y, height = z) %>%
mutate_at(.vars = c("length", "width", "height"), # Change this column
.funs = function(x) x / 10) %>% # Applying them the same function
filter(cut %in% c("Premium", "Ideal")) %>% # Filtering on the cut
select(-depth, -color, -table) # Exclude non wanted columnsThen, we use summarise() and a group_by!
dim_summary <- diamonds_modif %>%
group_by(cut, clarity) %>%
summarise_at(c("length", "width", "height"), mean)# A tibble: 16 x 5
# Groups: cut [2]
cut clarity length width height
<ord> <ord> <dbl> <dbl> <dbl>
1 Premium I1 0.684 0.679 0.413
2 Premium SI2 0.657 0.654 0.400
3 Premium SI1 0.605 0.601 0.369
4 Premium VS2 0.583 0.581 0.357
5 Premium VS1 0.574 0.571 0.351
6 Premium VVS2 0.539 0.537 0.330
7 Premium VVS1 0.504 0.502 0.308
8 Premium IF 0.523 0.522 0.319
9 Ideal I1 0.675 0.674 0.416
10 Ideal SI2 0.626 0.627 0.387
11 Ideal SI1 0.578 0.579 0.357
12 Ideal VS2 0.543 0.543 0.335
13 Ideal VS1 0.545 0.547 0.337
14 Ideal VVS2 0.521 0.523 0.322
15 Ideal VVS1 0.496 0.498 0.306
16 Ideal IF 0.483 0.485 0.2986 Manipulating multiple tables
6.1 Multiple tables
- Consider two different sources of data
- Don’t hesitate to run the code to see what they look like!
# If required
# install.packages("nycflights13")
# Data on New York city flights and airports
library(nycflights13) # New York city flights of 2013 data setError in library(nycflights13): aucun package nommé 'nycflights13' n'est trouvétable_1 <- airports %>%
select(name, lon, lat) %>% # Select those 3 columns
slice(1:2) # 2 first rowsError in select(., name, lon, lat): objet 'airports' introuvabletable_2 <- airports %>%
select(name, lon, lat) %>% # Same columns as table_1, different rows
slice(6:7)Error in select(., name, lon, lat): objet 'airports' introuvabletable_3 <- airports %>%
select(alt, tz) %>%
slice(1:2) # Same rows as table_1, different columnsError in select(., alt, tz): objet 'airports' introuvable6.2 Binding tables bind_rows() and bind_cols()
table_1 %>% # Binding table_2 below table_1
bind_rows(table_2) # Column names must match!Error in list2(...): objet 'table_1' introuvabletable_1 %>% # Binding table_3 next to table_1
bind_cols(table_3) # Number of rows must match!Error in list2(...): objet 'table_1' introuvable6.3 Joining tables
6.3.1 Left jointure left_join()
- Suppose we have the two following tables
drivers <- tibble(name = c("Sue", "Sue", "Marc", "Gunter", "Rayan", "Rayan"),
car = c("Clio", "ZX", "AX", "Lada", "Twingo", "Clio"))
vehicles <- tibble(car = c("Twingo", "Ferrari", "Clio", "Lada", "ZX"),
speed = c("140", "280", "160", "85", "160"))
drivers# A tibble: 6 x 2
name car
<chr> <chr>
1 Sue Clio
2 Sue ZX
3 Marc AX
4 Gunter Lada
5 Rayan Twingo
6 Rayan Clio vehicles# A tibble: 5 x 2
car speed
<chr> <chr>
1 Twingo 140
2 Ferrari 280
3 Clio 160
4 Lada 85
5 ZX 160 - We would like to joint the tables to obtain the speed for each driver
left_join(drivers, vehicles, by = "car")# A tibble: 6 x 3
name car speed
<chr> <chr> <chr>
1 Sue Clio 160
2 Sue ZX 160
3 Marc AX <NA>
4 Gunter Lada 85
5 Rayan Twingo 140
6 Rayan Clio 160 - As no record was given for the speed of an AX, an
NAis inserted. - The Ferrari car is omitted as it was not present in the first table
- Watch out:
left_join()is not symmetric
left_join(vehicles, drivers, by = "car")# A tibble: 6 x 3
car speed name
<chr> <chr> <chr>
1 Twingo 140 Rayan
2 Ferrari 280 <NA>
3 Clio 160 Sue
4 Clio 160 Rayan
5 Lada 85 Gunter
6 ZX 160 Sue left_join(x, y)is equivalent toright_join(y, x)
6.3.2 Inner jointures inner_join()
inner_join()only keeps lines where info is present on both tables
inner_jointure <- inner_join(drivers, vehicles, by = "car")- Both the
MarcandFerrarilines are missing.
6.3.3 Full jointures full_join()
full_join()keeps all lines, puttingNAif necessary
full_jointure <- full_join(drivers, vehicles, by = "car")- Both the
MarcandFerrarilines are present, withNA.
6.4 Exercise
For the diamonds_modif object, compute, for each combination cut/clarity, the number of observations per group. Include it in the table dim_summary created in the previous exercise (previous section).
6.5 Exercise (Solution)
diamonds_modif %>%
group_by(cut, clarity) %>%
summarise(Nb_obs = n()) %>%
left_join(dim_summary, by = c("cut", "clarity"))# A tibble: 16 x 6
# Groups: cut [2]
cut clarity Nb_obs length width height
<ord> <ord> <int> <dbl> <dbl> <dbl>
1 Premium I1 205 0.684 0.679 0.413
2 Premium SI2 2949 0.657 0.654 0.400
3 Premium SI1 3575 0.605 0.601 0.369
4 Premium VS2 3357 0.583 0.581 0.357
5 Premium VS1 1989 0.574 0.571 0.351
6 Premium VVS2 870 0.539 0.537 0.330
7 Premium VVS1 616 0.504 0.502 0.308
8 Premium IF 230 0.523 0.522 0.319
9 Ideal I1 146 0.675 0.674 0.416
10 Ideal SI2 2598 0.626 0.627 0.387
11 Ideal SI1 4282 0.578 0.579 0.357
12 Ideal VS2 5071 0.543 0.543 0.335
13 Ideal VS1 3589 0.545 0.547 0.337
14 Ideal VVS2 2606 0.521 0.523 0.322
15 Ideal VVS1 2047 0.496 0.498 0.306
16 Ideal IF 1212 0.483 0.485 0.2987 Cleaning and transforming data
When you have a dataset, there are three interrelated rules which make your dataset tidy:
Each variable must have its own column
Each observation must have its own row
Each value must have its own cell
This mean that you must have a clear idea of what is a variable, an observation, or even a value!
7.1 A toy example: grades dataset
grades <- tibble(
Name = c("Tommy", "Mary", "Gary", "Cathy"),
Sex_age = c("m_15", "f_15", "m_16", "f_14"),
Test1 = c(10, 15, 16, 14),
Test2 = c(11, 13, 10, 12),
Test3 = c(12, 13, 17, 10)
)| Name | Sex_age | Test1 | Test2 | Test3 |
|---|---|---|---|---|
| Tommy | m_15 | 10 | 11 | 12 |
| Mary | f_15 | 15 | 13 | 13 |
| Gary | m_16 | 16 | 10 | 17 |
| Cathy | f_14 | 14 | 12 | 10 |
One can spot multiple potential problems:
- The
Sex_agecolumn gathers two variables - There are three variables tests, which can be what we want, but we might want a variable
Gradewhose value is a grade (and might depend on which test was performed!)
7.2 Splitting a column into 2: separate()
- The
Sex_agecolumn should be two columns (two variables)
grades_s_a <- # grades with two columns sex age
grades %>%
separate(Sex_age, into = c("Sex", "Age"), sep = "_")
print(grades_s_a)# A tibble: 4 x 6
Name Sex Age Test1 Test2 Test3
<chr> <chr> <chr> <dbl> <dbl> <dbl>
1 Tommy m 15 10 11 12
2 Mary f 15 15 13 13
3 Gary m 16 16 10 17
4 Cathy f 14 14 12 10- The
separgument can handle regular expressions. - Not trivial but really powerful tool for string manipulation.
7.3 Merging two columns: unite()
- The inverse of separate is
unite()
grades_s_a %>%
unite(col = "Sex_age", Sex, Age, sep = "_") %>%
knitr::kable(format = "html")| Name | Sex_age | Test1 | Test2 | Test3 |
|---|---|---|---|---|
| Tommy | m_15 | 10 | 11 | 12 |
| Mary | f_15 | 15 | 13 | 13 |
| Gary | m_16 | 16 | 10 | 17 |
| Cathy | f_14 | 14 | 12 | 10 |
7.4 Gathering columns: pivot_longer()
- Suppose that the observation of interest is the grade
- In
grade_s_a, the a same variable (the grade) is coded in three columns. - We might gather them in one column, and keep test number in one column.
- The table will be a in long format that might be better suited for our use.
- We use the function
pivot_longer.
grades_long <- grades_s_a %>%
pivot_longer(starts_with("Test"), # Gathered columns
names_to = "Test", # Column gathering old column names
values_to = "Grade") # Column gathering values
grades_long# A tibble: 12 x 5
Name Sex Age Test Grade
<chr> <chr> <chr> <chr> <dbl>
1 Tommy m 15 Test1 10
2 Tommy m 15 Test2 11
3 Tommy m 15 Test3 12
4 Mary f 15 Test1 15
5 Mary f 15 Test2 13
6 Mary f 15 Test3 13
7 Gary m 16 Test1 16
8 Gary m 16 Test2 10
9 Gary m 16 Test3 17
10 Cathy f 14 Test1 14
11 Cathy f 14 Test2 12
12 Cathy f 14 Test3 10- This format is often better for visualisation with
ggplot. - The dimension of
grades_gatheredis now \(12\times 5\). - Alternative code not enumerating all columns to gather, but excluding non gathered ones
grades_long <- grades_s_a %>%
pivot_longer(cols = -c("Name", "Age", "Sex"), # non gathered columns
names_to = "Test", # Column gathering old column names
values_to = "Grade") # Column gathering values
grades_long# A tibble: 12 x 5
Name Sex Age Test Grade
<chr> <chr> <chr> <chr> <dbl>
1 Tommy m 15 Test1 10
2 Tommy m 15 Test2 11
3 Tommy m 15 Test3 12
4 Mary f 15 Test1 15
5 Mary f 15 Test2 13
6 Mary f 15 Test3 13
7 Gary m 16 Test1 16
8 Gary m 16 Test2 10
9 Gary m 16 Test3 17
10 Cathy f 14 Test1 14
11 Cathy f 14 Test2 12
12 Cathy f 14 Test3 107.4.1 Why longer format?
- The longer format is (often) more convenient for summarizing.
- If we want the mean grade per test, for each sex
grades_long %>% # Initial data
group_by(Sex, Test) %>% # Grouping of interest
summarise(Mean = mean(Grade)) # Get the summary of interest, per group# A tibble: 6 x 3
# Groups: Sex [2]
Sex Test Mean
<chr> <chr> <dbl>
1 f Test1 14.5
2 f Test2 12.5
3 f Test3 11.5
4 m Test1 13
5 m Test2 10.5
6 m Test3 14.57.5 Get a wider table pivot_wider()
- The inverse of
pivot_longerispivot_wider() - It spreads a column into multiples, according to a key column
# The sep argument is used to give correct names to column
grades_long %>%
pivot_wider(names_from = "Test", values_from = "Grade")# A tibble: 4 x 6
Name Sex Age Test1 Test2 Test3
<chr> <chr> <chr> <dbl> <dbl> <dbl>
1 Tommy m 15 10 11 12
2 Mary f 15 15 13 13
3 Gary m 16 16 10 17
4 Cathy f 14 14 12 107.6 Input missing data: complete()
- Sometimes, when a value is missing, it is not recorded
notes <- tibble(Nom = c("Alain", "Alain", "Benoit", "Claire"),
Discipline = c("Maths", "Francais", "Maths", "Francais"),
Note = c(16, 9, 17, 11),
Present = rep("oui", 4))| Nom | Discipline | Note | Present |
|---|---|---|---|
| Alain | Maths | 16 | oui |
| Alain | Francais | 9 | oui |
| Benoit | Maths | 17 | oui |
| Claire | Francais | 11 | oui |
- Benoit has no note in French , Claire has no note in maths.
- You might want to complete the data such that each Name/Discipline has a note.
notes %>%
complete(Nom, Discipline)# A tibble: 6 x 4
Nom Discipline Note Present
<chr> <chr> <dbl> <chr>
1 Alain Francais 9 oui
2 Alain Maths 16 oui
3 Benoit Francais NA <NA>
4 Benoit Maths 17 oui
5 Claire Francais 11 oui
6 Claire Maths NA <NA> - The default assignment is
NA. - You can assign any value to missing data
notes_completes <- notes %>%
complete(Nom, Discipline,
fill = list(Note = 0, Present = "non")) # Custom the value per column| Nom | Discipline | Note | Present |
|---|---|---|---|
| Alain | Francais | 9 | oui |
| Alain | Maths | 16 | oui |
| Benoit | Francais | 0 | non |
| Benoit | Maths | 17 | oui |
| Claire | Francais | 11 | oui |
| Claire | Maths | 0 | non |
7.7 Handling missing data: replace_na
Consider the following table which as NA
donnees_na <- tibble(Groupe = rep(c("A", "B"), each = 3),
Nom = c("Al", "Bob", NA, "Dave", "Elle", "Fanch"),
Note = c(NA, 8, 7, 4.5, 1, 4))
donnees_na# A tibble: 6 x 3
Groupe Nom Note
<chr> <chr> <dbl>
1 A Al NA
2 A Bob 8
3 A <NA> 7
4 B Dave 4.5
5 B Elle 1
6 B Fanch 4 - Suppose we want to assign values to
NA. - Use
replace_na.
donnees_na %>%
replace_na(replace = list(Nom ="Unknown",
Note = 0)) # Assign same value for all columns# A tibble: 6 x 3
Groupe Nom Note
<chr> <chr> <dbl>
1 A Al 0
2 A Bob 8
3 A Unknown 7
4 B Dave 4.5
5 B Elle 1
6 B Fanch 4 - This syntax might not be well suited if many columns are present
- By combining with
mutate_if, one can do several things
7.7.1 Replace NA based on a condition
# Replace NA only in numeric column
donnees_na %>%
mutate_if(is.numeric, # We change only if it's numeric
# The treatment is made column wise
function(colonne) replace_na(colonne, replace = 0))# A tibble: 6 x 3
Groupe Nom Note
<chr> <chr> <dbl>
1 A Al 0
2 A Bob 8
3 A <NA> 7
4 B Dave 4.5
5 B Elle 1
6 B Fanch 4 7.7.2 Replace NA by mean of the column
- It is standard to replace by the mean of the column
# Replace NA by the mean of the numeric column
donnees_na %>%
mutate_if(is.numeric, # We change only if it's numeric
# The treatment is made column wise
function(colonne) replace_na(colonne,
replace = mean(colonne, na.rm = TRUE)))# A tibble: 6 x 3
Groupe Nom Note
<chr> <chr> <dbl>
1 A Al 4.9
2 A Bob 8
3 A <NA> 7
4 B Dave 4.5
5 B Elle 1
6 B Fanch 4 - Of course, by combining with
group_by, you can replace theNAby the mean of the group (here, the group A)
# Replace NA by the group mean
donnees_na %>%
group_by(Groupe) %>% # We first group by the required group
# And then apply the same treatment
mutate_if(is.numeric, # We change only if it's numeric
# The treatment is made column wise
function(colonne) replace_na(colonne,
replace = mean(colonne, na.rm = TRUE)))# A tibble: 6 x 3
# Groups: Groupe [2]
Groupe Nom Note
<chr> <chr> <dbl>
1 A Al 7.5
2 A Bob 8
3 A <NA> 7
4 B Dave 4.5
5 B Elle 1
6 B Fanch 4 7.8 Exercise
From the diamonds_modif table of the previous exercise, create the diamonds_gathered table having the following the following column structure:
# A tibble: 5 x 5
price cut clarity Dimension Value
<int> <ord> <ord> <chr> <dbl>
1 10096 Premium SI1 width 0.731
2 655 Ideal VVS1 width 0.439
3 593 Premium SI2 height 0.264
4 765 Ideal VS1 width 0.443
5 3543 Ideal VS2 length 0.5737.9 Exercise (solution)
diamonds_gathered <- diamonds_modif %>%
select(price, cut, clarity, length, width, height) %>%
gather(key = "Dimension", value = "Value", length, width, height)8 Dealing with characters
8.1 stringr package
The stringr provides powerful functions to handle characters.
For instance, let’s consider the following data.
For a full cover of all these possibilities, there are are lot of dedicated pages, for instance, this one.
example <- tibble(Nom = c("Mr_Al_Bob", "Mr_Bob_Col"),
Age = c(42, 41))8.2 Replacing character pattern
- Say we want to replace
Mrby aM. - Use the
str_replacefunction
example %>%
mutate(Nom = str_replace(Nom, pattern = "Mr", replacement = "M."))# A tibble: 2 x 2
Nom Age
<chr> <dbl>
1 M._Al_Bob 42
2 M._Bob_Col 41- If we want to replace all the
_by a space
example %>%
mutate(Nom = str_replace_all(Nom, pattern = c("_"), replacement = c(" ")))# A tibble: 2 x 2
Nom Age
<chr> <dbl>
1 Mr Al Bob 42
2 Mr Bob Col 41- We can of course remove the
Mrin front, withstr_remove
example %>%
mutate(Nom = str_replace_all(Nom, pattern = c("_"), replacement = c(" ")),
Nom = str_remove(Nom, pattern = "Mr "))# A tibble: 2 x 2
Nom Age
<chr> <dbl>
1 Al Bob 42
2 Bob Col 419 Dealing with factors
9.1 Transforming numeric in factor
Let’s consider the following table
athletes <- tibble(Name = c("Alice", "Bob", "Charles", "Dan", "Elsa", "Fanch"),
Performance = c(15, 10, 5, 18, 11, 4),
Aptitude = c("Strong", "Intermediate", "Weak",
"Strong", "Intermediate", "Weak"),
Qualified = c(1, 1, 0, 1, 1, 0))
athletes# A tibble: 6 x 4
Name Performance Aptitude Qualified
<chr> <dbl> <chr> <dbl>
1 Alice 15 Strong 1
2 Bob 10 Intermediate 1
3 Charles 5 Weak 0
4 Dan 18 Strong 1
5 Elsa 11 Intermediate 1
6 Fanch 4 Weak 0The column Qualified is here considered a a numeric, whereas it’s a qualitative, 1 meaning “yes”, 0 meaning no.
This can pose problem either for graphical representation or, worse, for model fitting in R.
# We want a boxplot, with one box per level of Qualified
ggplot(athletes) +
aes(x = Qualified, y = Performance) +
geom_boxplot() # Do not work, as Qualified is coded as numeric!Warning: Continuous x aesthetic -- did you forget aes(group=...)?
To change it, we will change (using mutate) the column to factor.
athletes <- athletes %>%
mutate(Qualified = factor(Qualified)) # Changed to qualified
ggplot(athletes) +
aes(x = Qualified, y = Performance) +
geom_boxplot() # Now it works
9.2 Changing factor labelling
The factor Qualified now has two levels, which are 0 and 1. In the graph above, it would be way better if 0/1 was understandable by a user. We can change it using the labels argument to relabel the levels. Watch out, the labels must be specified in the same order than the levels
athletes %>%
mutate(Qualified = factor(Qualified,
levels = c(0, 1), # Specify the level order
labels = c("No", "Yes"))) %>% # And rename in the same order
ggplot() +
aes(x = Qualified, y = Performance) +
geom_boxplot() # Now it works
9.3 Change characters in factor
For plots or model, there is no absolute need to transform characters in factor, as R understands that it is qualitative attribute.
# A boxplot of performance per Aptitude
ggplot(athletes) +
aes(x = Aptitude, y = Performance) +
geom_boxplot()
Here, the order of the levels is given by alphabetical order, and is not natural. We then can transform Aptitude to factor to specify the right levels order.
athletes %>%
mutate(Aptitude = factor(Aptitude,
# Now specify the levels order
# Caution, the levels given must appear in the data!
# f you want to change labels, use labels = ...
levels = c("Weak", "Intermediate", "Strong"))) %>%
# A boxplot of performance per Aptitude
ggplot() +
aes(x = Aptitude, y = Performance) +
geom_boxplot() # The levels are now in natural order