How to encode a factor variable as numeric in R

March 14, 2026 · 2 min read · Updated March 14, 2026 · beginner

r factors data-types tidyverse

Factors store categorical data with discrete levels. Sometimes you need the underlying numeric codes instead of the character labels.

The basics: as.numeric()

When you convert a factor directly to numeric, you get the internal integer codes:

# Create a factor
education <- factor(c("high school", "bachelors", "masters", "phd"))
education
# [1] high school bachelors masters    phd   
# Levels: bachelors high school masters phdd

# Convert to numeric - this gives you the codes!
as.numeric(education)
# [1] 2 1 3 4

This works, but the codes are alphabetical by default (1 = bachelors, 2 = high school).

Get numeric codes in your level order

If you defined a specific order, extract those codes properly:

# Define custom level order
education <- factor(
  c("high school", "bachelors", "masters", "phd"),
  levels = c("high school", "bachelors", "masters", "phd")
)

# Now the codes match your order
as.numeric(education)
# [1] 1 2 3 4

Extract numeric from factor levels

Need the level positions without the internal encoding? Use this pattern:

# Method 1: Use as.numeric() on the levels, then index
x <- factor(c("low", "medium", "high"))
as.numeric(x)
# [1] 2 3 1

# Method 2: Match against levels explicitly
x <- factor(c("low", "medium", "high"))
match(x, levels(x))
# [1] 2 3 1

Using dplyr and tidyr

In the tidyverse, create a numeric column with mutate():

library(dplyr)

df <- data.frame(
  id = 1:5,
  category = factor(c("low", "medium", "high", "low", "high"))
)

df <- df %>%
  mutate(category_num = as.numeric(category))

df
#   id category category_num
# 1  1      low            1
# 2  2   medium            2
# 3  3     high            3
# 4  4      low            1
# 5  5      high            3

Ordered factors

Ordered factors preserve the order but still encode as integers:

# Create ordered factor
satisfaction <- factor(
  c("low", "medium", "high", "low"),
  levels = c("low", "medium", "high"),
  ordered = TRUE
)

# Numeric conversion respects the order
as.numeric(satisfaction)
# [1] 1 2 3 1

# You can also use as.integer()
as.integer(satisfaction)
# [1] 1 2 3 1

Preserve original numeric values from character

If your factor was originally numeric, you might want the original numbers:

# When factor was created from numbers
rating <- factor(c(1, 5, 3, 4, 2))

# Get the original numbers back
as.numeric(as.character(rating))
# [1] 1 5 3 4 2

This matters because as.numeric(factor(c(1, 5, 3))) gives you c(1, 3, 2) (the codes), not c(1, 5, 3) (the original values).