How to Create Categorical Variables in R

Categorical variables are crucial in statistical modeling and data analysis, representing distinct categories or levels that are mutually exclusive. In R, categorical variables are usually stored as factors. This comprehensive guide will walk through several methods to create and manipulate categorical variables, enhancing your data analysis toolkit in R.

1. Basics of Categorical Variables

Categorical variables can be either nominal, representing different categories without intrinsic order, or ordinal, depicting categories with a meaningful sequence. Creating accurate categorical variables is paramount for meaningful analysis and correct statistical inference.

2. Creating Categorical Variables with factor( )

The fundamental way to create categorical variables in R is by using the factor() function.

2.1. Basic Usage of factor( )

# Creating a basic factor
colors <- c("Red", "Blue", "Green")
colors_factor <- factor(colors)

2.2. Ordered Factors

To create an ordered factor, specify the ordered argument as TRUE and specify the levels argument to define the order of the levels.

# Creating an ordered factor
temperatures <- c("High", "Low", "Medium")
temperatures_factor <- factor(temperatures, ordered = TRUE, levels = c("Low", "Medium", "High"))

3. Using gl( ) Function to Generate Factors

The gl() function is especially helpful when you need to create factor levels by specifying the pattern of repetition.

# Generating a factor with two levels, each repeated twice
factor_var <- gl(2, 2, labels = c("Male", "Female"))

4. Creating Categorical Variables within Data Frames

Often, categorical variables are created and manipulated within data frames.

4.1. Defining Factors within Data Frames

# Creating a data frame with a factor
df <- data.frame(
  ID = 1:3,
  Color = factor(c("Red", "Blue", "Green"))
)

4.2. Converting Character Columns to Factors

By default, character columns in a data frame are converted to factors. However, you can control this behavior using the stringsAsFactors argument.

# Creating a data frame without automatic conversion to factors
df <- data.frame(
  ID = 1:3,
  Color = c("Red", "Blue", "Green"),
  stringsAsFactors = FALSE
)

5. Manipulating Levels of Categorical Variables

Managing the levels of categorical variables is crucial for proper data analysis and visualization.

5.1. Changing the Order of Levels

# Reordering levels
colors_factor <- factor(colors_factor, levels = c("Green", "Red", "Blue"))

5.2. Dropping Unused Levels

Using the droplevels() function, you can remove unused levels from a factor.

Suppose we have a factor with several levels, and we create a subset of data that doesn’t include all the original levels. In such cases, it is useful to drop the unused levels using droplevels().

Here’s a step-by-step example:

# Define a factor
fruits_factor <- factor(c("Apple", "Banana", "Cherry", "Grapes"))

# Create a subsetted factor that does not include all levels of the original factor
subsetted_factor <- fruits_factor[fruits_factor %in% c("Apple", "Banana")]

# Displaying the subsetted_factor reveals that it still retains the levels “Cherry” and “Grapes”, even though they are not present in the data.
print(subsetted_factor)
# [1] Apple  Banana
# Levels: Apple Banana Cherry Grapes

# Dropping unused levels
trimmed_factor <- droplevels(subsetted_factor)

# Displaying the trimmed_factor will show that it only has the levels that are present in the data.
print(trimmed_factor)
# [1] Apple  Banana
# Levels: Apple Banana

Conclusion

Creating and managing categorical variables effectively are foundational skills in statistical computing with R. By understanding the methods to create factors and manipulate their levels, analysts can construct meaningful representations of categorical data. Whether it is creating ordered factors, generating factors with specific patterns using gl(), or managing factor levels within data frames, each technique serves as a stepping stone to more accurate and insightful data analysis.