Chapter 6 More about Factors

This was part of a impromptu session learning about factors.

# load libraries
library(tidyverse)
library(readxl)
library(janitor)
library(forcats)

smoke_complete <- read_excel("data/smoke_complete.xlsx", 
                             sheet=1, 
                             na="NA")

6.1 Making a factor variable out of disease

We’re adding a fourth value, BRCA to our levels here.

smoke_complete2 <- smoke_complete %>% 
    mutate(disease_factor = 
               factor(disease,
                      levels = c("LUSC", "CESC", "BLCA", "BRCA")
                      )
           ) 

6.2 Using the character variable

ggplot(smoke_complete2) + 
  aes(x=disease, y=cigarettes_per_day) +
  geom_boxplot()

6.3 Compare to the factor variable

ggplot(smoke_complete2) + aes(x=disease_factor, y=cigarettes_per_day) +
  geom_boxplot()

6.4 Another thing about factors

Factor levels also specify the permissible values.

In this example, LUSC and BRCA are the permissible values. We pass a character vector into them, and you can see those values (BLCA, CESC) are recoded as NAs

character_vector <- c("LUSC", "LUSC", "BRCA", "BLCA", "BRCA", "CESC", "CESC")

factor_vector <- factor(character_vector, levels=c("LUSC", "BRCA"))

factor_vector

## [1] LUSC LUSC BRCA <NA> BRCA <NA> <NA>
## Levels: LUSC BRCA

6.5 fct_rev() - reversing the order of a factor

Very useful when using factors on the y-axis, because the default ordering is first value at the bottom, rather than first value at the top.

library(forcats)

#fct_rev()

smoke_complete3 <- smoke_complete2 %>%
  mutate(disease_rev = fct_rev(disease_factor))

#show original factor
ggplot(smoke_complete3) + aes(y=disease_factor, x=cigarettes_per_day) +
  geom_boxplot()

#show factor with reversed order
ggplot(smoke_complete3) + aes(y=disease_rev, x=cigarettes_per_day) +
  geom_boxplot()

6.6 fct_reorder()

fct_reorder() lets you reorder factors by another numeric variable.

library(forcats)

#fct_rev()

smoke_complete3 <- smoke_complete2 %>%
  mutate(disease_reorder = fct_reorder(disease_factor, cigarettes_per_day))

ggplot(smoke_complete3) + aes(y=disease_reorder, x=cigarettes_per_day) +
  geom_jitter()

6.7 fct_collapse

fct_collapse() lets you collapse multiple categories into one category.

smoke_complete3 %>% 
  mutate(disease_collapse = fct_collapse(
    disease_factor,
    other = c("BLCA", "CESC"),
    LUSC = c("LUSC")
  )) %>%
  tabyl(disease_collapse)

##  disease_collapse   n   percent
##              LUSC 836 0.7256944
##             other 316 0.2743056
##              BRCA   0 0.0000000

6.8 Other really useful forcats functions

fct_recode() - lets you recode values manually.

fct_other() - lets you define what categories are in an other variable.