csv - How to do histograms of this row-column table in R ggplot?

Question

I am trying to plot the descriptive variables in the first row by the following procedure. I also tried unsuccessfully with quoting the column/row names

1. rotate rows and columns in the CSV data for the correposding data structure (tall table) required in the thread A very simple histogram with R? with ggplot

2. to plot histogram of events as Absolute variable XOR (Average, Min, Max)

   • If absolute value only, just draw absolute value in histogram.
   • If (average, min and max), just draw them in the histogram with whiskers (= whisker plot) where the limits of the whiskers are made by the min and max.

Data

1. initially, data.csv

   "Vars"    , "Sleep", "Awake", "REM", "Deep"
   "Absolute",        ,       , 5     , 7
   "Average" , 7      , 12    ,       ,
   "Min"     , 4      , 5     ,       , 
   "Max"     , 10     , 15    ,       ,
   

2. data after reshaping visually

               V1       V2       V3       V4
   Vars  Absolute Average  Min      Max     
   Sleep     <NA>        7        4       10
   Awake     <NA>       12        5       15
   REM          5     <NA>     <NA>     <NA>
   Deep         7     <NA>     <NA>     <NA>
   

3. data after reshaping for R

    data <- structure(list(V1 = structure(c(3L, NA, NA, 1L, 2L), .Names = c("Vars", 
    "Sleep", "Awake", "REM", "Deep"), .Label = c(" 5", " 7", "Absolute"
    ), class = "factor"), V2 = structure(c(3L, 2L, 1L, NA, NA), .Names = c("Vars", 
    "Sleep", "Awake", "REM", "Deep"), .Label = c("12", " 7", "Average "
    ), class = "factor"), V3 = structure(c(3L, 1L, 2L, NA, NA), .Names = c("Vars", 
   "Sleep", "Awake", "REM", "Deep"), .Label = c(" 4", " 5", "Min     "
    ), class = "factor"), V4 = structure(c(3L, 1L, 2L, NA, NA), .Names = c("Vars", 
   "Sleep", "Awake", "REM", "Deep"), .Label = c("10", "15", "Max     "
    ), class = "factor")), .Names = c("V1", "V2", "V3", "V4"), row.names = c("Vars", 
   "Sleep", "Awake", "REM", "Deep"), class = "data.frame")
   

R code with debugging code

dat.m <- read.csv("data.csv")

# rotate rows and columns
dat.m <- as.data.frame(t(dat.m)) # https://stackoverflow.com/a/7342329/54964 Comment 42-

library("reshape2")
dat.m <- melt(dat.m, id.vars="Vars")

## Just plot values existing there correspondingly    
library("ggplot2")
# https://stackoverflow.com/a/25584792/54964
# TODO following
#ggplot(dat.m, aes(x = "Vars", y = value,fill=variable)) 

Error

Error: id variables not found in data: Vars
Execution halted

R: 3.3.3, 3.4.0 (backports)
OS: Debian 8.7
R reshape2, ggplot2, ... with sessionInfo() after loading the two packages

Platform: x86_64-pc-linux-gnu (64-bit)

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C   

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] ggplot2_2.1.0  reshape2_1.4.2

loaded via a namespace (and not attached):
 [1] colorspace_1.3-2 scales_0.4.1     magrittr_1.5     plyr_1.8.4      
 [5] tools_3.3.3      gtable_0.2.0     Rcpp_0.12.10     stringi_1.1.5   
 [9] grid_3.3.3       stringr_1.2.0    munsell_0.4.3    

Testing HaberdashPI's proposal

Output in Fig. 1 where wrongly absolute value in Sleep and Awake. If NA, just set value to zero.

Fig. 1 HaberdashPI's proposal output not as expected

Data structure of dat.m before the transpose

'data.frame':   4 obs. of  5 variables:
 $ Absolute: Factor w/ 2 levels " 5"," 7": NA NA 1 2
  ..- attr(*, "names")= chr  "Sleep" "Awake" "REM" "Deep"
 $ Average : Factor w/ 2 levels "12"," 7": 2 1 NA NA
  ..- attr(*, "names")= chr  "Sleep" "Awake" "REM" "Deep"
 $ Min     : Factor w/ 2 levels " 4"," 5": 1 2 NA NA
  ..- attr(*, "names")= chr  "Sleep" "Awake" "REM" "Deep"
 $ Max     : Factor w/ 2 levels "10","15": 1 2 NA NA
  ..- attr(*, "names")= chr  "Sleep" "Awake" "REM" "Deep"
 $ Vars    : chr  "Sleep" "Awake" "REM" "Deep"
      Absolute Average  Min      Max       Vars
Sleep     <NA>        7        4       10 Sleep
Awake     <NA>       12        5       15 Awake
REM          5     <NA>     <NA>     <NA>   REM
Deep         7     <NA>     <NA>     <NA>  Deep

Data structure of dat.m after the transpose

'data.frame':   16 obs. of  3 variables:
 $ Vars    : chr  "Sleep" "Awake" "REM" "Deep" ...
 $ variable: Factor w/ 4 levels "Absolute","Average ",..: 1 1 1 1 2 2 2 2 3 3 ...
 $ value   : chr  NA NA " 5" " 7" ...

    Vars variable value
1  Sleep Absolute  <NA>
2  Awake Absolute  <NA>
3    REM Absolute     5
4   Deep Absolute     7
5  Sleep Average      7
6  Awake Average     12
7    REM Average   <NA>
8   Deep Average   <NA>
9  Sleep Min          4
10 Awake Min          5
11   REM Min       <NA>
12  Deep Min       <NA>
13 Sleep Max         10
14 Awake Max         15
15   REM Max       <NA>
16  Deep Max       <NA>

Testing akash87's proposal

Code

ds <- dat.m
str(ds)
ds
ds$variable
ds$variable %in% c("Min","Max")

Wrong output because all False in the end

 $ Vars    : chr  "Sleep" "Awake" "REM" "Deep" ...
 $ variable: Factor w/ 4 levels "Absolute","Average ",..: 1 1 1 1 2 2 2 2 3 3 ...
 $ value   : chr  NA NA " 5" " 7" ...
    Vars variable value
1  Sleep Absolute  <NA>
2  Awake Absolute  <NA>
3    REM Absolute     5
4   Deep Absolute     7
5  Sleep Average      7
6  Awake Average     12
7    REM Average   <NA>
8   Deep Average   <NA>
9  Sleep Min          4
10 Awake Min          5
11   REM Min       <NA>
12  Deep Min       <NA>
13 Sleep Max         10
14 Awake Max         15
15   REM Max       <NA>
16  Deep Max       <NA>
[1] "hello 3"
 [1] Absolute Absolute Absolute Absolute Average  Average  Average  Average 
 [9] Min      Min      Min      Min      Max      Max      Max      Max     
Levels: Absolute Average  Min      Max     
 [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
[13] FALSE FALSE FALSE FALSE

So doing ds[ds$variable %in% c("Min","Max"), ] will given False output because error-carried-forward.

Testing Uwe's proposal

Code with explicit data.table::dcast and two times data.table::melt. Printing out sessionInfo() just before molten <- .... Note library(ggplot2) is not loaded yet because the error comes from the line molten <- ....

$ Rscript test111.r 
    Vars "Average" "Max" "Min" Absolute
1: Sleep         7    10     4       NA
2: Awake        12    15     5       NA
3:   REM        NA    NA    NA        5
4:  Deep        NA    NA    NA        7
R version 3.4.0 (2017-04-21)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Debian GNU/Linux 8 (jessie)

Matrix products: default
BLAS: /usr/lib/openblas-base/libblas.so.3
LAPACK: /usr/lib/libopenblasp-r0.2.12.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  base     

other attached packages:
[1] data.table_1.10.4

loaded via a namespace (and not attached):
[1] compiler_3.4.0 methods_3.4.0 
Error in melt.data.table(transposed, measure.vars = c("Absolute", "Average")) : 
  One or more values in 'measure.vars' is invalid.
Calls: <Anonymous> -> melt.data.table
Execution halted

Testing Uwe's proposal with test code 2

Code

molten <- structure(list(Vars = structure(c(1L, 2L, 1L, 2L, 1L, 2L), class = "factor", .Label = c("V1", "V2")), variable = structure(c(1L, 1L, 2L, 2L, 3L, 3L), class = "factor", .Label = c("ave", "ave_max", "lepo")), value = c(7L, 8L, 10L, 10L, 4L, 4L)), .Names = c("Vars", "variable", "value"), row.names = c(NA, -6L), class = c("data.table", "data.frame"))

print(molten)

library(ggplot2)
ggplot(molten, aes(x = Vars, y = value, fill = variable, ymin = lepo, ymax = ave_max)) + 
  geom_col() + geom_errorbar(width = 0.2)

Output

  Vars variable value
1   V1      ave     7
2   V2      ave     8
3   V1  ave_max    10
4   V2  ave_max    10
5   V1     lepo     4
6   V2     lepo     4
Error in FUN(X[[i]], ...) : object 'lepo' not found
Calls: <Anonymous> ... by_layer -> f -> <Anonymous> -> f -> lapply -> FUN -> FUN
Execution halted

Answer 1

The problem with your code is that you used "Vars" with a quote instead of simple Vars in the ggplot aes function. Also, the header of your data set is messed up. The Absolute, Average, ... should be the column names of the data set, not the values themselves. That's why you get the error from melt function.

Given your data set, here is my attempt:

#Data
data = cbind.data.frame(c("Sleep", "Awake", "REM", "Deep"),
                        c(NA, NA, 5, 7),
                        c(7, 12, NA, NA),
                        c(4, 5, NA, NA),
                        c(10, 15, NA, NA))
colnames(data) = c("Vars", "Absolute", "Average", "Min", "Max")

#reshape
dat.m <- melt(data, id.vars="Vars")
#Stacked plot
ggplot(dat.m, aes(x = Vars, y = value)) + geom_bar(aes(fill=variable), stat = "identity")

This will produce:

#Or multiple bars
ggplot(dat.m, aes(x = Vars, y = value)) + 
  geom_bar(aes(fill=variable), stat = "identity", position="dodge") 

#Or separated by Vars
ggplot(dat.m, aes(x = Vars, y = value)) + geom_bar(aes(fill=variable), stat = "identity", position="dodge") + facet_wrap( ~ Vars, scales="free")

I am adding another graph to the answer. This collaborates @Uwe answer.

#data
data <- structure(list(Vars = structure(1:2, class = "factor", .Label = c("V1", "V2")), ave = c(7L, 8L), ave_max = c(10L, 10L), lepo = c(4L, 4L)), .Names = c("Vars", "ave", "ave_max", "lepo"), row.names = c(NA, -2L), class = c("data.table", "data.frame"), sorted = "Vars")
#Melt
library(data.table)
mo = data.table::melt(data, measure.vars = c("ave"))
ggplot(mo, aes(x = Vars, y = value, fill = variable, ymin = lepo, ymax = ave_max)) + geom_col() + geom_errorbar(width = 0.2)

This will produce: