Table of categorical variables by a grouping variable in R

问题

I have a dataset with some categorical variables + a "cluster" variable. For example:

time <- c("Morning", "Evening" ,"Morning", "Morning", "Afternoon", "Evening", "Afternoon")
dollar <- c("1-5", "6-10", "11-15", "1-5", "1-5", "6-10", "6-10")
with_kids <- c("no", "yes", "yes", "no", "no", "yes", "yes")
cluster <- c(1,1,2,3,2,2,3)

data <- cbind(time, dollar, with_kids, cluster)

How can I create a frequency table of all the categorical variables by "cluster"?

Desired output is the table on the right (column % of each categorical variable within each cluster).

I know this code will work for one variable. What is the most efficient way to do it if I have many more categorical variables?

table(data$time, data$cluster)

回答1:

I'm not entirely sure of your desired output, but here are two possibilities.

A list of tables:

myList <- lapply(dat[head(names(dat), -1)], table, dat$cluster)
myList
$time

            1 2 3
  Afternoon 0 1 1
  Evening   1 1 0
  Morning   1 1 1

$dollar

        1 2 3
  1-5   1 1 1
  11-15 0 1 0
  6-10  1 1 1

$with_kids

      1 2 3
  no  1 1 1
  yes 1 2 1

To get a list of proportion tables, you can lapply your list of tables using prop.table as the function and feed it margin=2:

lapply(myList, prop.table, margin=2)
$time

                    1         2         3
  Afternoon 0.0000000 0.3333333 0.5000000
  Evening   0.5000000 0.3333333 0.0000000
  Morning   0.5000000 0.3333333 0.5000000

$dollar

                1         2         3
  1-5   0.5000000 0.3333333 0.5000000
  11-15 0.0000000 0.3333333 0.0000000
  6-10  0.5000000 0.3333333 0.5000000

$with_kids

              1         2         3
  no  0.5000000 0.3333333 0.5000000
  yes 0.5000000 0.6666667 0.5000000

to rbind them together

do.call(rbind, lapply(dat[head(names(dat), -1)], table, dat$cluster))
          1 2 3
Afternoon 0 1 1
Evening   1 1 0
Morning   1 1 1
1-5       1 1 1
11-15     0 1 0
6-10      1 1 1
no        1 1 1
yes       1 2 1

data

dat <- 
structure(list(time = structure(c(3L, 2L, 3L, 3L, 1L, 2L, 1L), .Label = c("Afternoon", 
"Evening", "Morning"), class = "factor"), dollar = structure(c(1L, 
3L, 2L, 1L, 1L, 3L, 3L), .Label = c("1-5", "11-15", "6-10"), class = "factor"), 
    with_kids = structure(c(1L, 2L, 2L, 1L, 1L, 2L, 2L), .Label = c("no", 
    "yes"), class = "factor"), cluster = c(1, 1, 2, 3, 2, 2, 
    3)), .Names = c("time", "dollar", "with_kids", "cluster"), row.names = c(NA, 
-7L), class = "data.frame")

回答2:

time <- c("Morning", "Evening" ,"Morning", "Morning", "Afternoon", "Evening", "Afternoon")
dollar <- c("1-5", "6-10", "11-15", "1-5", "1-5", "6-10", "6-10")
with_kids <- c("no", "yes", "yes", "no", "no", "yes", "yes")
cluster <- c(1,1,2,3,2,2,3)
data <- data.frame(time, dollar, with_kids, cluster)

You can use the dplyr package and select as many variables as you like

library(dplyr)
data %>% 
  group_by(interaction(time, cluster, dollar)) %>% 
  summarise(count = n())

# A tibble: 7 x 2
  `interaction(time, cluster, dollar)` count
                                <fctr> <int>
1                        Morning.1.1-5     1
2                      Afternoon.2.1-5     1
3                        Morning.3.1-5     1
4                      Morning.2.11-15     1
5                       Evening.1.6-10     1
6                       Evening.2.6-10     1
7                     Afternoon.3.6-10     1

来源：https://stackoverflow.com/questions/45354436/table-of-categorical-variables-by-a-grouping-variable-in-r

标签

frequency