Agrupando níveis de fatores

Dados Categóricos no Tidyverse

Emily Robinson

Data Scientist

Um gráfico de barras, intitulado "frequência de alturas na pesquisa de etiqueta em voos da 538", com altura no eixo x e contagem no eixo y. O eixo x vai de "menos de 5 ft" a "6'6" ou mais", com uma barra para cada polegada.

fct_collapse()

flying_etiquette %>%
  mutate(height = fct_collapse(height, 
             under_5_3 = c("Under 5 ft.", "5'0\"", "5'1\"", "5'2\""),
             over_6_1 = c("6'1\"", "6'2\"", "6'3\"", "6'4\"", 
             "6'5\"", "6'6\" and above"))) %>%
  pull(height) %>%
  levels()

 [1] "under_5_3" "5'10\""    "5'11\""    "5'3\""    
 [5] "5'4\""     "5'5\""     "5'6\""     "5'7\""    
 [9] "5'8\""     "5'9\""     "6'0\""     "over_6_1"

fct_other(): keep

flying_etiquette %>%
  mutate(new_height = fct_other(height, keep = c("6'4\"", "5'1\""))) %>%
  count(new_height)

# A tibble: 4 x 2
  new_height     n
  <fct>      <int>
1 "5'1\""       19
2 "6'4\""       11
3 Other        828
4 NA           182

fct_other(): drop

flying_etiquette %>%
  mutate(new_height = fct_other(height, 
  drop = c("Under 5 ft.", "5'0\"", "5'1\"", "5'2\"", "5'3\""))) %>%
  pull(new_height) %>%
  levels()

 [1] "5'4\""           "5'5\""           "5'6\""          
 [4] "5'7\""           "5'8\""           "5'9\""          
 [7] "5'10\""          "5'11\""          "6'0\""          
[10] "6'1\""           "6'2\""           "6'3\""          
[13] "6'4\""           "6'5\""           "6'6\" and above"
[16] "Other"

fct_lump_prop()

flying_etiquette %>%
  mutate(new_height = fct_lump_prop(height, prop = .08)) %>%
  count(new_height)

  new_height     n
  <fct>      <int>
1 "5'4\""       79
2 "5'6\""       75
3 "5'7\""       76
4 "5'8\""       76
5 Other        552
6 NA           182

fct_lump_n()

flying_etiquette %>%
  mutate(new_height = fct_lump_n(height, n = 3)) %>%
  count(new_height)

  new_height     n
  <fct>      <int>
1 "5'4\""       79
2 "5'7\""       76
3 "5'8\""       76
4 Other        627
5 NA           182

Vamos praticar!

Dados Categóricos no Tidyverse