Code Handout - R for Social Scientists

This document contains all of the core functions that were covered in the R for Social Scientists workshop. Each function is presented alongside an example of how it can be used. It is split into the following 4 sections:

• Introduction to R

• Starting with Data

• Data Wrangling

• Data Visualization

Each section has instructions to load all necessary libraries or data, so you can start from any of the 4 points linked above.

Introduction to R

The first section covers core programming concepts and functions in Base R and does not require any data or libraries to be loaded.

Creating Objects

• <- – “assignment arrow”, assigns a value (vector, dataframe, single value) to the name of a variable

R

x <- 3
• c() – the “concatenate” function combines inputs to form a vector, the values have to be the same data type.

R

animals <- c("bird", "cat", "dog")
numbers <- c(1, 14, 57, 89)
logicals <- c(TRUE, FALSE, TRUE, TRUE)
• + – addition and other mathematical operators can be repeated on every value in a vector.

R

y <- c(1, 2, 3)
z <- x + y

Inspecting Objects

• str() – compact display of the structure of an R object

R

str(animals)
• class() – returns the type of element of any R object

R

class(logicals)
• typeof() – returns the data type or storage mode of any R object

R

typeof(numbers)

Functions in R

• args() – returns the arguments of a function

R

args(round)
• named arguments – the name of the argument the function expects
• You can choose to not name your arguments, if you know the exact order they should be in!
• However, we generally discourage this.
• round() – round a decimal or fraction to a specified number of digits

R

# Either of these work, since the digits argument is named explicitly.
round(3.14159, digits = 2)
round(digits = 2, 3.14159)

# This does not work, since the arguments are not named and in the incorrect order.
round(2, 3.14159)

Functions to Summarize Data

• sqrt() – returns the square root of a numeric variable

R

sqrt(numbers)
• mean() – returns the mean of a numeric variable
• You can add the na.rm argument, to remove NA values before calculating the mean.

R

mean(numbers)
• max() – returns the maximum of a numeric variable
• You can add the na.rm argument, to remove NA values before calculating the max.

R

max(numbers)
• sum() – returns the sum of a numeric variable
• You can add the na.rm argument, to remove NA values before calculating the sum.

R

sum(numbers)
• length() – returns the length of a vector (of any datatype)

R

length(animals)
• var() – find the variance of a numerical variable
• sd() – finds the standard deviation of a numerical variable
• IQR() – find the innerquartile range (Q3 - Q1) of a numerical variable
• median() – finds the median of a numerical variable

Subsetting Data

• [] – used to subset elements from a vector

• X:Y – used to retrieve a “slice” of a vector starting at X and continuing through Y

R

animals[3]
# selects the third element

animals[2:3]
# selects the second and third element

animals[c(1, 3)]
# selects the first and third element
• relational operators – return logical values indicating where a relation is satisfied. The most commonly used logical operators for data analysis are as follows:
• == means “equal to”
• != means “not equal to”
• > or < means “greater than” or “less than”
• >= or <= means “greater than or equal to” or “less than or equal to”

R

animals == "dog"

animals != "cat"

numbers > 4

numbers <= 12
• logical operators – join subset criteria together
• & means “and” – where two criteria must both be satisfied
• | means “or” – where at least one criteria must be satisfied

R

numbers > 4 & numbers < 20

animals == "dog" | animals == "cat"
• %in% – the “inclusion operator”, allows you to test if any of the elements of a search vector (on the left hand side) are found in the target vector (on the right hand side).
• The levels of the target vector must be included in a vector (c()).

R

possessions <- c("car", "bicycle", "radio", "television", "mobile_phone")

possessions %in% c("car", "bicycle", "motorcycle")

Missing Data

• is.na() – returns a vector of logical values indicating which elements of a vector have NA values
• Often combined with !, where the ! negates the previous statement (e.g. !TRUE is equal to FALSE).

R

missing <- c(1, 3, NA, 7, 12, NA)

is.na(missing)

!is.na(missing)
• na.omit() – removes the observations with NA values

R

na.omit(missing)
• complete.cases() – returns a vector of logical values indicating which elements of a vector are not missing (NA) values

R

complete.cases(missing)

Starting with Data

In this section, we begin working with data. All data examples are in the context of the Palmer Penguins, found here (link).

Packages

Packages (also called libraries) expand the capabilities of R beyond the functions that come when you install it. Each needs to be downloaded and installed only once but loaded into each R session.

• install.packages() – install a new package

• library() – loads packages into your R session

R

# Install packages (not run)
# Delete # from lines below if missing packages
#install.packages("tidyverse")
#install.packages("lubridate")
#install.packages("palmerpenguins")

library(tidyverse)
library(lubridate)
library(palmerpenguins) #load Palmer Penguins data as penguins

Inspecting Data

• dim() - returns a vector with the number of rows as the first element, and the number of columns as the second element (the dimensions of the object)

R

dim(penguins)
• nrow() - returns the number of rows
• ncol() - returns the number of columns

R

nrow(penguins)
ncol(penguins)
• head() - displays the first 6 rows of the dataframe
• tail() - displays the last 6 rows of the dataframe

R

head(penguins)
tail(penguins)
• names() - returns the all of the names of an object (both row and column)
• colnames() - returns column names for dataframes (without row names)

R

names(penguins)
colnames(penguins)
• glimpse() - provides a preview of the data, where column names are presented with their associated data types, and the entries from each column are printed in each row

R

glimpse(penguins)
• str() - returns the structure of the object and information about the class, the names and data types of each column, and a preview of the first entries of each column

R

str(penguins)
• summary() - provides summary statistics for each column
• Note: summary statistics for character variables are not meaningful, as they simply state the number of observations (length) of the variable

R

summary(penguins)

Subsetting Data in Dataframes

• [] – selects rows and columns from a dataframe
• The first entry is the row number, the second entry is the column number(s), and they are separated with a comma.

R

# Selects the element in the first row, second column
penguins[1, 2]

# Selects every element in the fourth row
penguins[4, ]

# Selects every element in the third column
penguins[, 3]
• [[]] – selects a column from a dataframe
• Inside the brackets you can pass either the number of the column or the name of the column (in quotations)

R

penguins[[1]]

penguins[["island"]]

Working with Different Data Types

• factor() – creates a categorical variable from a character or numeric variable, variable has a factor datatype
• the values (level) of the factor levels is specified in the levels argument, where the levels must be specified in a vector (using c())
• Note: the order you wish for the levels to appear is how you should list them in the levels argument, you can also specify ordered = TRUE to ensure the levels remain in this order

R

penguins$year_fct <- factor(penguins$year,
levels = c("2007", "2008", "2009"),
ordered = TRUE)
• as.factor() – creates a categorical variable from a character or numeric variable, variable has a factor datatype
• does not allow for you to specify the order of the levels
• defaults to alphabetical ordering for factor levels

R

penguins$year_fct <- as.factor(penguins$year)
• levels() – returns the levels of a factor variable in the order they were stored
• Note: this function will not work for character variables

levels(penguins$year_fct) • nlevels() – returns the number of levels of a factor variable • Note: this function will not work for character variables R nlevels(penguins$year_fct)
• as.character() – creates a character variable from a numeric or factor variable

R

penguins$species_chr <- as.character(penguins$species)
• ymd() – transforms dates stored as character or numeric variables to dates
• Note: to use this function, dates must be stored in year-month-day format
• The function does well with heterogeneous formats (as seen below), but formats where some of the entries are not in double digits may not be parsed correctly.

R

x <- c("2009-01-01", "2009-01-02", "2009-01-03")
ymd(x)
• day() – extracts the day (number) of a date variable

R

day(x)
• month() – extracts the month (number) of a date variable

R

month(x)
• year() – extracts the year of a date variable

R

year(x)

Basic Data Visualization (see Data Visualization section for more)

• plot() – a generic function for plotting R objects
• In this lesson plot() was used to create bargraphs of categorical variables.

R

plot(penguins\$species)

Data Wrangling

This section continues using the Palmer Penguins data and introduces concepts and functions to explore, clean and summarize data, many of which come from the dplyr and plyr tidyverse libraries.

Packages

• library() – loads packages into your R session

R

library(tidyverse)
library(palmerpenguins)

Inspecting Data

• glimpse() – shows a summary of the dataset, the number of rows and columns, variable names, and the first 10 entries of each variable

R

glimpse(penguins)

Verbs of Data Wrangling

• %>% – the “pipe” operator, joins sequences of data wrangling steps together, works with any function that has data = as the first argument
• select() – selects variables (columns) from a dataframe

R

penguins %>%
select(species)
• filter() – filters observations (rows) out of / into a dataframe, where the inputs (arguments) are the conditions to be satisfied in the data that are kept

Logical operators: Filtering for certain observations (e.g. flights from a particular airport) is often of interest in data frames where we might want to examine observations with certain characteristics separately from the rest of the data. To do so, you can use the filter function and a series of logical operators. The most commonly used logical operators for data analysis are as follows:

• == means “equal to”

• != means “not equal to”

• > or < means “greater than” or “less than”

• >= or <= means “greater than or equal to” or “less than or equal to”

R

# It's nice to have a new line for each condition, so your code is easier to read!
penguins %>%
body_mass_g > 3000,
year == 2008)
• mutate() – creates new variables or modifies existing variables

R

penguins %>%
filter(is.na(bill_length_mm) != TRUE,
is.na(bill_depth_mm) != TRUE) %>%
mutate(body_mass_kg = body_mass_g / 1000)
• group_by() – groups the dataframe based on levels of a categorical variable, usually used alongside summarize()

R

penguins %>%
group_by(island)
• summarize()-- creates data summaries of variables in a dataframe, for grouped summaries use alongsidegroup_by()

R

penguins %>%
filter(is.na(body_mass_g) != TRUE) %>%
group_by(island) %>%
summarize(mean_mass = mean(body_mass_g))  
• ungroup() – removes the grouping of a dataframe, typically used after group summaries when additional ungrouped operations are required

R

penguins %>%
filter(is.na(body_mass_g) != TRUE) %>%
group_by(island) %>%
summarize(mean_mass = mean(body_mass_g)) %>%
ungroup() 
• arrange() – orders a dataframe based on the values of a numerical variable, paired with desc() to order in descending order

R

penguins %>%
filter(is.na(body_mass_g) != TRUE) %>%
group_by(island) %>%
summarize(mean_mass = mean(body_mass_g)) %>%
arrange(desc(mean_mass))

Chain multiple operations together with %>% to create specific outputs without extra steps or dataframes being created along the way.

R

penguins %>%
select(species, island, body_mass_g, sex, year) %>%
filter(island ==   "Torgersen",
is.na(body_mass_g) != TRUE) %>%
group_by(species, year) %>%
summarize(mean_mass = mean(body_mass_g),
median_mass = median(body_mass_g),
observations = n()) %>%
arrange(desc(mean_mass))

Other Data Wrangling Tools

• count() – counts the number of observations (rows) of the different levels of a categorical variable
• can add sort = TRUE to sort the table in descending order (similar to using arrange(desc()) )

R

penguins %>%
count(species)
• sample_n() – selects $n$ rows from the dataframe, based on the value of size specified

R

penguins %>%
sample_n(size = 10)
• replace_na() – replaces NA values with the value specified
• The values to be replaced must be passed to the function (input) as a list() object.

R

penguins %>%
replace_na(list(bill_length_mm = "no_measurement",
bill_depth_mm = "no_measurement")) %>%
glimpse()
• separate_rows() – separates a variable with multiple values based on the delimiter specified.

• Variables whose entries are stored as a list with commas or semicolons are great candidates for this function!
• rowSums() – forms row sums for numeric variables

• Note: In the lesson rowSums() was used on a logical variable, because logical values can be numerically represented as 0 (FALSE) and 1 (TRUE)

R

x <- tibble(x1 = 3, x2 = c(4:1, 2:5))
rowSums(x)

Pivoting Dataframes

• pivot_wider() – transforms a dataframe from long to wide format
• takes three principal arguments:
1. the data (often passed by a %>%)
2. the names_from column variable whose values will become new column names
3. the values_from column variable whose values will fill the new column variables.
• Further arguments include values_fill which, if set, fills in missing values with the value provided.

R

wide <- penguins %>%
mutate(island_logical = TRUE) %>%
pivot_wider(names_from = species,
values_from = island_logical,
values_fill = list(island_logical = FALSE))

glimpse(wide)
• pivot_longer() – transforms a dataframe from wide to long format
• takes four principal arguments:
1. the data
2. cols are the names of the columns we use to fill the a new values variable (or to drop).
3. the names_to column variable we wish to create from the cols provided.
4. the values_to column variable we wish to create and fill with values associated with the cols provided.

R

wide %>%
names_to = "species",
values_to = "island_logical")

Extracting Data

• write_csv() – writes a dataframe to a csv file, output into the file path specified

R

write_csv(wide, path = "data/penguins_wide.csv")

Importing Data

• read_csv() – function to import a csv file.
• First argument is the path to the data, passed as a character (inside quotations).
• You can specify what values should be considered missing, using the na argument.

R

penguins_wide <- read_csv("data/penguins_wide.csv")

Data Visualization with ggplot2

This section continues using the Palmer Penguins data to introduce key features of the ggplot2 package for visualizing data and provide examples combining data wrangling and visualization.

R

library(tidyverse)
library(palmerpenguins)

Foundations of ggplot()

• ggplot() – a function to create the shell of a visualization, where specific variables are mapped to different aspects of the plot

• aes() – aesthetics that can be used when creating a ggplot(), where the aesthetics can either be hard coded (e.g. color = "blue") or associated with a variable (e.g. color = sex).

• The following are the aesthetic options for most plots:
• x – variable to use for x axis
• y – variable to use for y axis
• alpha – changes transparency
• color – produces colored outline
• fill – fills with color
• group – used with categorical variables, similar to color

R

#nothing should appear on this plot except the axes and labels
penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species))
• + – an important aspect creating a ggplot() is to note that the geom_XXX() function is separated from the ggplot() function with a plus sign, +.

• ggplot() plots are constructed in series of layers, where the plus sign separates these layers.
• Generally, the + sign can be thought of as the end of a line, so you should always hit enter/return after it. While it is not mandatory to move to the next line for each layer, doing so makes the code a lot easier to organize and read.
• geom_point() – adds a scatter plot; see full explanation later in list

R

penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
geom_point()

Geometric Objects to Visualize the Data

• geom_histogram() – adds a histogram to the plot, where the observations are binned into ranges of values and then frequencies of observations are plotted on the y-axis
• You can specify the number of bins you want with the bins argument

R

penguins %>%
ggplot(aes(x = bill_length_mm)) +
geom_histogram(bins = 20)
• geom_boxplot() – adds a boxplot to the plot, where observations are aggregated (summarized), the min, Q1, median, Q3, and maximum are plotted as the box and whiskers, and “outliers” are plotted as points.
• You can plot a vertical boxplot by specifying the x variable, or a horizontal boxplot by specifying the y variable.
• Note: the min and max may not be included in the whiskers, if they are deemed to be “outliers” based on the $1.5 \\times \\text{IQR}$ rule.

R

# Horizontal boxplot
penguins %>%
ggplot(aes(x = bill_length_mm)) +
geom_boxplot()

# Vertical boxplot
penguins %>%
ggplot(aes(y = bill_length_mm)) +
geom_boxplot()
• geom_density() – adds a density curve to the plot, where the probability density is plotted on the y-axis (so the density curve has a total area of one).
• By default this creates a density curve without shading. By specifying a color in the fill argument, the density curve is shaded.
• Can be thought of as the “one group” violin plot (see below)

R

penguins %>%
ggplot(aes(x = bill_length_mm)) +
geom_density(fill = "tomato")
• geom_violin() – plots violins for each level of a categorical variable
• Can be thought of as a hybrid mix of geom_boxplot() and geom_density(), as the density is displayed, but it is reflected to provide a plot similar in nature to a boxplot.
• To obtain violins stacked vertically, declare the categorical variable as y. To obtain side-by-side violins, declare the categorical variable as x.

R

# Stacked vertically
penguins %>%
ggplot(aes(x = bill_length_mm, y = species)) +
geom_violin()

# Side-by-side
penguins %>%
ggplot(aes(y = bill_length_mm, x = species)) +
geom_violin()
• geom_bar() – creates a barchart of a categorical variable
• Can produce stacked barcharts by specifying a variable as the fill aesthetic.
• Can change from stacked barchart to a side-by-side barchart by specifying position = "dodge".
• If your data are already in counts (e.g. output from count()), then you can specify the stat = "identity" argument inside geom_bar().

R

# Stacked barchart
penguins %>%
ggplot(aes(x = species)) +
geom_bar(aes(fill = sex))

# Side-by-side barchart
penguins %>%
ggplot(aes(x = species)) +
geom_bar(aes(fill = sex),
position = "dodge")

# If data are raw counts
penguins %>%
count(species, sex) %>%
ggplot(aes(x = species, y = n)) +
geom_bar(aes(fill = sex),
stat = "identity",
position = "dodge")
• geom_point() – plots each observation as an (x, y) point, used to create scatterplots
• Can use alpha to increase the transparency of the points, to reduce overplotting.
• Can specify aesthetics inside of geom_point() for local aesthetics (point level) or inside of ggplot() for global aesthetics (plot level)

R

penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm)) +
geom_point(aes(color = species))
• geom_jitter() – plots each observation as an (x, y) point and adds a small amount of jitter around the point
• Useful so that we can see each point in the locations where there are overlapping points.
• Can specify the width and height of the jittering using the optional arguments.

R

penguins %>%
ggplot(aes(y = body_mass_g, x = species)) +
geom_violin() +
geom_jitter(aes(color = sex), width = 0.25, height = 0.25)
• geom_smooth() – plots a line over a set of points, draws the readers eye to a specific trend
• The methods we will use are “lm” for a linear model (straight line), and “loess” for a wiggly line
• By default, the smoother gives you gray SE bars, to remove these add se = FALSE

R

penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
geom_point() +
geom_smooth(method = "lm") 
• facet_wrap() – creates subplots of your original plot, based on the levels of the variable you input
• To facet by one variable, use ~variable.
• To facet by two variables, use variable1 ~ variable2.
• If you prefer for your facets to be organized in rows or columns, use the nrow and/or ncol arguments.

R

penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
geom_point() +
geom_smooth(method = "lm") +
facet_wrap(~island, nrow = 1)

Plot Characteristics

• labs() – specifies the plot labels, possible labels are: x, y, color, fill, title, and subtitle

R

penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
geom_point() +
geom_smooth(method = "lm") +
labs(x = "Bill Length (mm)",
y = "Bill Depth (mm)",
color = "Penguin Species")
• theme_bw() – changes the plotting background to the classic dark-on-light ggplot2 theme.
• This theme may work better for presentations displayed with a projector.
• Other common themes are theme_minimal(), theme_light(), and theme_void().

R

penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
geom_point() +
geom_smooth(method = "lm") +
labs(x = "Bill Length (mm)",
y = "Bill Depth (mm)",
color = "Penguin Species") +
theme_bw()
• theme() – adjust individual theme elements
• Possible options are:
• panel.grid – controls the grid lines (panel.grid = element_blank() removes grid lines)
• text – specifies font size for the entire plot (e.g. text = element_text(size = 16)
• axis.text.x – specifies the font size for the x-axis text
• axis.text.y – specifies the font size for the y-axis text
• plot.title – specifies aspects of the plot title, can use plot.title = element_text(hjust = 0.5) to centre the title

R

penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
geom_point() +
geom_smooth(method = "lm") +
labs(x = "Bill Length (mm)",
y = "Bill Depth (mm)",
color = "Penguin Species") +
theme_bw() +
theme(axis.text.x = element_text(size = 12),
axis.text.y = element_text(size = 12))

Exporting Plots

• ggsave() – convenient function for saving a plot
• Unless specified, defaults to the last plot that was made.
• Uses the size of the current graphics device to determine the size of the plot.

R

plot1 <- penguins %>%
ggplot(aes(x = bill_length_mm, y = bill_depth_mm, color = species)) +
geom_point() +
geom_smooth(method = "lm") +
facet_wrap(~island, nrow = 1)

ggsave(path = "images/faceted_plot.png", plot = plot1)`