0

I'm analysing some data, and I've come a difficulty - I don't know how to summarise my whole dataset using the variable names as the "group".

A sample data:

structure(list(x4 = c(3, 4, 7, 4, 5, 1, 5, 2, 7, 1), x5 = c(2, 
4, 4, 4, 5, 3, 6, 1, 7, 1), x6 = c(3, 5, 4, 7, 5, 4, 6, 4, 6, 
2), x7 = c(4, 1, 6, 4, 6, 4, 6, 2, 7, 2), x9 = c(5, 5, 4, 5, 
6, 3, 7, 5, 6, 1), x10 = c(3, 6, 5, 4, 6, 5, 6, 3, 6, 1), x11 = c(6, 
7, 7, 7, 6, 7, 7, 5, 7, 4), x12 = c(6, 7, 6, 7, 6, 4, 6, 6, 7, 
5), x14 = c(5, 7, 5, 6, 4, 6, 6, 5, 6, 4), x15 = c(4, 7, 7, 7, 
6, 4, 6, 5, 6, 1), x16 = c(4, 7, 7, 7, 6, 5, 7, 3, 6, 4), x17 = c(4, 
5, 5, 7, 6, 6, 7, 4, 6, 2), x18 = c(3, 4, 7, 7, 6, 5, 6, 4, 6, 
2), x19 = c(5, 7, 5, 7, 6, 6, 6, 3, 6, 1), x22 = c(4, 4, 5, 7, 
6, 7, 6, 5, 6, 2), x26 = c(6, 7, 5, 4, 6, 7, 7, 4, 6, 4), x29 = c(4, 
7, 2, 7, 6, 4, 7, 3, 6, 1), x33 = c(3, 7, 7, 7, 6, 5, 6, 3, 6, 
3), x34 = c(5, 5, 4, 7, 6, 7, 7, 5, 6, 2), x35 = c(4, 4, 7, 7, 
5, 7, 6, 4, 6, 2), x36 = c(4, 7, 6, 7, 6, 5, 5, 4, 6, 2), x37 = c(3, 
4, 7, 4, 5, 4, 6, 3, 5, 2), x49 = c(4, 7, 7, 7, 6, 5, 5, 6, 6, 
3), x50 = c(4, 7, 6, 5, 5, 5, 6, 5, 7, 4)), row.names = c(NA, 
-10L), class = "data.frame", .Names = c("x4", "x5", "x6", "x7", 
"x9", "x10", "x11", "x12", "x14", "x15", "x16", "x17", "x18", 
"x19", "x22", "x26", "x29", "x33", "x34", "x35", "x36", "x37", 
"x49", "x50"))

I just want some statistics, like this:

summary <- dados_afc %>% 
  summarise_all(funs(mean, sd, mode, median))

But the result is a df with one observation and lots of variable. I wanted it to have 5 columns: varname, mean, sd, mode, median, but I'm not sure how to do it. Any tips?

marc_s
  • 732,580
  • 175
  • 1,330
  • 1,459
GVianaF
  • 59
  • 7

1 Answers1

2

Note: I am not aware of a built-in way to get mode from R. See here for some discussion: Is there a built-in function for finding the mode?

# From the top answer there:
Mode <- function(x) {
  ux <- unique(x)
  ux[which.max(tabulate(match(x, ux)))]
}

To treat each column as a group, you can use tidyr::gather to convert your "wide" data into long form, and then dplyr::group_by to create groups with their own summary calculations:

library(tidyverse)  
summary <- dados_afc %>%
  gather(group, value) %>%
  group_by(group) %>%
  summarise_all(funs(mean, sd, Mode, median))


>   summary
# A tibble: 24 x 5
   group  mean    sd    Mode median
   <chr> <dbl> <dbl>   <dbl>  <dbl>
 1 x10     4.5 1.72        6    5  
 2 x11     6.3 1.06        7    7  
 3 x12     6   0.943       6    6  
 4 x14     5.4 0.966       6    5.5
 5 x15     5.3 1.89        7    6  
 6 x16     5.6 1.51        7    6  
 7 x17     5.2 1.55        6    5.5
 8 x18     5   1.70        6    5.5
 9 x19     5.2 1.87        6    6  
10 x22     5.2 1.55        6    5.5
Jon Spring
  • 55,165
  • 4
  • 35
  • 53