How can I summarizing data statistics using R

Question

how can I write a short script that creates a new data frame that reports the following descriptive statistics for each column of continuous data for the survey below: mean,

Answer 1

The following use of do.call, rbind and sapply provides a summary for each column that has the class 'numeric'. You can write your own statistics function if you need different statistics than those of summary (see the answer of @Jilber).

mtcars$carb = as.factor(mtcars$carb)  # Forcing one column to a factor
do.call('rbind', sapply(mtcars, function(x) if(is.numeric(x)) summary(x)))
       Min. 1st Qu.  Median     Mean 3rd Qu.    Max.
mpg  10.400  15.420  19.200  20.0900   22.80  33.900
cyl   4.000   4.000   6.000   6.1880    8.00   8.000
disp 71.100 120.800 196.300 230.7000  326.00 472.000
hp   52.000  96.500 123.000 146.7000  180.00 335.000
drat  2.760   3.080   3.695   3.5970    3.92   4.930
wt    1.513   2.581   3.325   3.2170    3.61   5.424
qsec 14.500  16.890  17.710  17.8500   18.90  22.900
vs    0.000   0.000   0.000   0.4375    1.00   1.000
am    0.000   0.000   0.000   0.4062    1.00   1.000
gear  3.000   3.000   4.000   3.6880    4.00   5.000

Answer 2

Here are some examples using data.table. I'm using the functions defined in the previous answers.

my.summary <- function(x, na.rm=TRUE){
  result <- c(Mean=mean(x, na.rm=na.rm),
              SD=sd(x, na.rm=na.rm),
              Median=median(x, na.rm=na.rm),
              Min=min(x, na.rm=na.rm),
              Max=max(x, na.rm=na.rm), 
              N=length(x))
}
set.seed(123)

df <- data.frame(id = 1:1000,
                 Distance = rnorm(1000, 50, 100),
                 Age = rnorm(1000, 50, 100),
                 Height = rnorm(1000, 50, 100)
                 )
df$Coning <- as.factor(ifelse(df$Distance > 0, "Yes", "No"))
library(fBasics)
library(data.table)
DT <- data.table(df)
setkey(DT, id)

Group by factor variable "Coning"

DT[,lapply(.SD,my.summary),by="Coning"]

Using my.summary() and basicStats() Just numeric Variables

DT[,lapply(.SD, my.summary),, .SDcols = names(DT)[2:4]]

BS <- DT[,sapply(.SD, basicStats),, .SDcols = names(DT)[2:4]]
BS[, summary := znames]
setnames(BS, 1:3, names(DT)[2:4])
BS

DT[,lapply(.SD, summary),, .SDcols = names(DT)[2:4]]

using summary() Numeric Variable using

DT[,sapply(.SD, function(x) if(is.numeric(x)) summary(x)),, .SDcols = names(DT)[2:4]]

Factor Variable

DT[,sapply(.SD, function(x) if(is.factor(x)) summary(x)),, .SDcols = names(DT)[5]]

Using the quantile function is also quite useful:

DT[,sapply(.SD, function(x) if(is.numeric(x)) quantile(x)),, .SDcols = names(DT)[2:4]]

Answer 3

You can write your own function to get such a summary into a data.frame:

# Defining the function
my.summary <- function(x, na.rm=TRUE){
  result <- c(Mean=mean(x, na.rm=na.rm),
              SD=sd(x, na.rm=na.rm),
              Median=median(x, na.rm=na.rm),
              Min=min(x, na.rm=na.rm),
              Max=max(x, na.rm=na.rm), 
              N=length(x))
}

# identifying numeric columns
ind <- sapply(df, is.numeric)


# applying the function to numeric columns only
sapply(df[, ind], my.summary)  
        Distance       Age     Height
Mean    58.67200 11.840000  1.9160000
SD      45.48137  4.604168  0.9796626
Median  48.80000 13.500000  1.7000000
Min      8.70000  4.000000  0.6000000
Max    241.80000 19.000000  5.0000000
N       50.00000 50.000000 50.0000000

Or you can use the built-in function basicStats from fBasics package for a more detailed summary:

> library(fBasics)
> basicStats(df[, ind])
               Distance        Age    Height
nobs          50.000000  50.000000 50.000000
NAs            0.000000   0.000000  0.000000
Minimum        8.700000   4.000000  0.600000
Maximum      241.800000  19.000000  5.000000
1. Quartile   28.300000   7.000000  1.125000
3. Quartile   74.675000  15.750000  2.475000
Mean          58.672000  11.840000  1.916000
Median        48.800000  13.500000  1.700000
Sum         2933.600000 592.000000 95.800000
SE Mean        6.432037   0.651128  0.138545
LCL Mean      45.746337  10.531510  1.637583
UCL Mean      71.597663  13.148490  2.194417
Variance    2068.555118  21.198367  0.959739
Stdev         45.481371   4.604168  0.979663
Skewness       1.711028  -0.158853  0.905415
Kurtosis       3.753948  -1.574527  0.578684