6 Factors

6.1 Factors

A factor is a data type similar to a vector. However, the values contained in a factor can only be selected from a set of levels.

houses_vector <- c("Bungalow", "Flat", "Flat",
  "Detached", "Flat", "Terrace", "Terrace")
houses_vector

## [1] "Bungalow" "Flat"     "Flat"     "Detached" "Flat"     "Terrace" 
## [7] "Terrace"

houses_factor <- factor(c("Bungalow", "Flat", "Flat",
  "Detached", "Flat", "Terrace", "Terrace"))
houses_factor

## [1] Bungalow Flat     Flat     Detached Flat     Terrace  Terrace 
## Levels: Bungalow Detached Flat Terrace

6.2 table

The function table can be used to obtain a tabulated count for each level.

houses_factor <- factor(c("Bungalow", "Flat", "Flat",
  "Detached", "Flat", "Terrace", "Terrace"))
houses_factor

## [1] Bungalow Flat     Flat     Detached Flat     Terrace  Terrace 
## Levels: Bungalow Detached Flat Terrace

table(houses_factor)

## houses_factor
## Bungalow Detached     Flat  Terrace 
##        1        1        3        2

6.3 Specified levels

A specific set of levels can be specified when creating a factor by providing a levels argument.

houses_factor_spec <- factor(
  c("People Carrier", "Flat", "Flat", "Hatchback",
      "Flat", "Terrace", "Terrace"),
  levels = c("Bungalow", "Flat", "Detached",
       "Semi", "Terrace"))

table(houses_factor_spec)

## houses_factor_spec
## Bungalow     Flat Detached     Semi  Terrace 
##        0        3        0        0        2

6.4 (Unordered) Factors

In statistics terminology, (unordered) factors are categorical (i.e., binary or nominal) variables. Levels are not ordered.

income_nominal <- factor(
  c("High", "High", "Low", "Low", "Low",
      "Medium", "Low", "Medium"),
  levels = c("Low", "Medium", "High"))
income_nominal > "Low"

## Warning in Ops.factor(income_nominal, "Low"): '>' not meaningful for
## factors

## [1] NA NA NA NA NA NA NA NA

6.5 Ordered Factors

In statistics terminology, ordered factors are ordinal variables. Levels are ordered.

income_ordered <- ordered(
  c("High", "High", "Low", "Low", "Low",
      "Medium", "Low", "Medium"),
  levels = c("Low", "Medium", "High"))
income_ordered > "Low"

## [1]  TRUE  TRUE FALSE FALSE FALSE  TRUE FALSE  TRUE

sort(income_ordered)

## [1] Low    Low    Low    Low    Medium Medium High   High  
## Levels: Low < Medium < High