HOME DATA MANIPULATION IN R TABLE R

Data Manipulation in R Data transformation

The `table`

function can be used for summarizing categorical data and generating absolute frequency and contigency tables. In this tutorial we will be exploring its syntax, various arguments, and practical examples to illustrate its utility in analyzing data. We will also explore the `prop.table`

for relative frequency tables, `xtabs`

for cross-tabulation, and `addmargins`

to add margins to tables.

## The `table`

function

The `table`

function in R is used to tabulate categorical data, counting the number of occurrences of each category. This function can create one-way tables, which provide the frequency of each category in a single variable, and two-way tables (or higher in high-dimensional arrays), which display the frequency distribution across two or more variables.

### Syntax of the `table`

function

The syntax of the function is the following:

`table(..., exclude = if (useNA == "no") c(NA, NaN), useNA = c("no", "ifany", "always"), dnn = list.names(...), deparse.level = 1)`

Being:

`...`

: one or more categorical variables or expressions to be tabulated.`exclude`

: optional argument specifying levels to be excluded from the table.`useNA`

: treatment of missing values. Possible options are`"no"`

(default),`"ifany"`

and`"always"`

`dnn`

: character vector providing names for the resulting table.`deparse.level`

: controls how`dnn`

is constructed by default. Read the documentation of the function for additional details.

### One-way frequency tables

One-way tables represent the frequency distribution of a single variable. They are useful for understanding the distribution of categories within a variable. In order to create a table you will need to input a character vector, as illustrated below:

`# Sample datadata <- c("B", "A", "C", "C", "A", "C", "B")# Create a simple frequency table for a categorical variableone_way_table <- table(data)one_way_table`

`dataA B C 2 2 3 `

The previous ouput means that there are two elements that correspond to category `"A"`

, two correspond to `"B"`

and three to `"C"`

. The previous table can be represented using a bar plot:

`# Sample datadata <- c("B", "A", "C", "C", "A", "C", "B")# Create a simple frequency table for a categorical variableone_way_table <- table(data)# Plot the tablebarplot(one_way_table, col = 2:4, ylab = "Count")`

The function provides an argument named `exclude`

that can be used to **exclude some categories from the output table**. In the following example we are excluding `"B"`

.

`# Sample datadata <- c("B", "A", "C", "C", "A", "C", "B")# Exclude specific levels from the frequency tabletable(data, exclude = "B")`

`dataA C 2 3`

Sometimes, analyzing the **presence of missing values** is as important as the available data. The `useNA`

argument can be leveraged to include `NA`

values in the table. When set to `"ifany"`

the table will also count the number of missing values.

`# Sample datadata <- c("B", "A", NA, NA, "A", "C", "B")# Count NA values if anytable(data, useNA = "ifany")`

`data A B C <NA> 2 2 1 2 `

When set to `"always"`

the table will display the number of `NA`

values even if there were none. This is very useful for data checking.

`# Sample datadata <- c("B", "A", "A", "C", "B")# Count NA values even if there are nonetable(data, useNA = "always")`

`data A B C <NA> 2 2 1 0 `

### Two-way contingency tables

Two-way tables show the relationship between two categorical variables. They are crucial for examining the interactions between variables. This type of table can also be created with the `table`

function, but you will need to input two character vectors of the same length instead of one, as illustrated below.

`# Sample datagender <- c("Male", "Female", "Male", "Female", "Male")age_group <- c("Junior", "Senior", "Senior", "Junior", "Junior")# Create a two-way tabletwo_way_table <- table(gender, age_group)# View the tabletwo_way_table`

` age_groupgender Junior Senior Female 1 1 Male 2 1`

The previous data can be represented with a bar plot making use of the `barplot`

function or any other similar function:

`# Sample datagender <- c("Male", "Female", "Male", "Female", "Male", "Female")age_group <- c("Junior", "Senior", "Senior", "Junior", "Junior", "Senior")# Create a two-way tabletwo_way_table <- table(gender, age_group)# Plot the tablebarplot(two_way_table, col = 2:3, beside = TRUE, ylab = "Count")legend("topright", legend = c("Female", "Male"), fill = 2:3)`

## The `prop.table`

function

The `prop.table`

function takes a table created with `table`

and converts it into a **relative frequency table**, also known as proportion table.

`# Sample datagender <- c("Male", "Female", "Male", "Female", "Male")age_group <- c("Junior", "Senior", "Senior", "Junior", "Junior")# Create a two-way tabletwo_way_table <- table(gender, age_group)# Relative frequency tableprop.table(two_way_table)`

` age_groupgender Junior Senior Female 0.2 0.2 Male 0.4 0.2`

The function includes an argument named `margin`

. Setting `margin`

to `1`

calculates proportions based on the sum of each row, whereas setting it to `2`

calculates proportions based on the sum of each column.

`# Sample datagender <- c("Male", "Female", "Male", "Female", "Male")age_group <- c("Junior", "Senior", "Senior", "Junior", "Junior")# Create a two-way tabletwo_way_table <- table(gender, age_group)# Relative frequency tableprop.table(two_way_table, margin = 1)`

` age_groupgender Junior Senior Female 0.5000000 0.5000000 Male 0.6666667 0.3333333`

## The `xtabs`

function

A function related to `table`

is `xtabs`

. The `xtabs`

function allows creating contingency tables and it is **specially useful for grouped data and when working with data frames**. Unlike `table`

, it uses a formula syntax, which allows for more complex specifications and is ideal for statistical analysis.

`# Sample data framedf <- data.frame(x = c("G1", "G2", "G2", "G1", "G1", "G2"), y = c("A", "B", "B", "C", "A", "C"))# Contingency table with xtabstab <- xtabs(~ x + y, data = df)tab`

` yx A B C G1 2 0 1 G2 0 2 1`

An interesting feature of `xtabs`

is that it can **create weighted contingency tables**. The following example illustrates how to input weights using the column `w`

:

`# Sample data framedf <- data.frame(x = c("G1", "G2", "G2", "G1", "G1", "G2"), y = c("A", "B", "B", "C", "A", "C"), w = c(0.1, 0.2, 0.2, 0.1, 0.1, 0.3))# Weighted contingency tabletab <- xtabs(w ~ x + y, data = df)tab`

` yx A B C G1 0.2 0.0 0.1 G2 0.0 0.4 0.3`

## The `addmargins`

function

The `addmargins`

function in R is used to add row and/or column margins, usually representing sums or totals of the rows and/or columns to tables created with `table`

or similar functions. The syntax of the function is the following:

`addmargins(A, margin = NULL, FUN = sum, quiet = FALSE)`

Being:

`A`

: the input table.`margin`

: the desired margin. By default, the function calculates all margins, but when is set to 1, only row margins are calculated, and when set to 2, only column margins are calculated.`FUN`

: function to be applied to calculate the margins. It sums by default.`quiet`

: logical. If set to`TRUE`

suppress messages.

When the function is applied to a table both margins will be added by default, counting the number of elements for rows and columns.

`# Sample datagender <- c("Male", "Female", "Male", "Female", "Male")age_group <- c("Junior", "Senior", "Senior", "Junior", "Junior")# Create a two-way tabletwo_way_table <- table(gender, age_group)# Add marginsaddmargins(two_way_table)`

` age_groupgender Junior Senior Sum Female 1 1 2 Male 2 1 3 Sum 3 2 5`

However, if you only want to calculate the margins for rows or for columns you will need to set the `margin`

argument to `1`

or `2`

depending on your needs.

`# Sample datagender <- c("Male", "Female", "Male", "Female", "Male")age_group <- c("Junior", "Senior", "Senior", "Junior", "Junior")# Create a two-way tabletwo_way_table <- table(gender, age_group)# Add marginsaddmargins(two_way_table, margin = 2)`

` age_groupgender Junior Senior Sum Female 1 1 2 Male 2 1 3`

R version 4.3.2 (2023-10-31 ucrt)