Skip to content

Help please

10 messages · Rui Barradas, Jim Lemon, Richard M. Heiberger +2 more

Original
Areti Panopoulou
# 
Hello,

I am trying to make a stacked barplot with different behaviours
("inactive", "active", "foraging" etc) on different substrates ("tree",
"ground" etc). I have found this function:

# Stacked Bar Plot with Colors and Legend
counts <- table(mtcars$vs, mtcars$gear)
barplot(counts, main="Car Distribution by Gears and VS",
  xlab="Number of Gears", col=c("darkblue","red"),
  legend = rownames(counts))

But don't know how to apply it. Can anybody help me apply the function to
work with my variables ( I can send more information if necessary, or make
any clarifications).

This is a great help, thanks a lot.
Richard M. Heiberger
# 
install.packages("HH")
library(HH)
?likert

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Rui Barradas
    

    

      
      #
    

  


  

      
Hello,

Please read the posting guide at the end of this and every R-Help mail.
You should post the output of

dput(data)

or, if the data set 'data' is too big, the output of

dput(head(data, 20))

for us to be able to help you.

Rui Barradas

?s 18:29 de 10/03/21, Areti Panopoulou escreveu:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Richard M. Heiberger
    

    

      
      #
    

  


  

          

      
      
      
    

        
vs
gear  0  1
   3 12  3
   4  2 10
   5  4  1

            

      
      
      
    

        
If this isn't enough, ask a more specific question.

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Jim Lemon
    

    

      
      #
    

  


  

      
Hi Areti,
Maybe this will help:

scrounging<-data.frame(
 behav=sample(c("inactive","active","foraging","snoozing"),50,TRUE),
 substr=sample(c("tree","ground","vine","air"),50,TRUE))
scrounge.tab<-table(scrounging)
barplot(scrounge.tab)
legend(3.8,14,c("inactive","active","foraging","snoozing"),
 fill=c("gray80","gray60","gray40","gray20"))

Jim

On Thu, Mar 11, 2021 at 9:54 AM Areti Panopoulou

        
<aretipanopoulou at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Richard M. Heiberger
    

    

      
      #
    

  


  

          

      
      
      
    

        
substr
behav      air ground tree vine
  active     1      5    3    6
  foraging   2      6    3    0
  inactive   6      1    1    3
  snoozing   2      1    3    7

            

      
      
      
    

        

        

      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Rasmus Liland
    

    

      
      #
    

  


  

      
Dear Areti,

this dcast data.frame presents the same 
info as Jim's barplot

	reshape2::dcast(data=scrounging, formula=behav~substr, fun.aggregate=length)

yielding

	     behav air ground tree vine
	1   active   5      4    3    3
	2 foraging   2      1    3    4
	3 inactive   0      4    4    5
	4 snoozing   3      5    1    3

I tried to use likert by the example in 
the example in ?likert::likert

	library(likert)
	data(pisaitems)
	items29 <- pisaitems[,substr(names(pisaitems), 1,5) == 'ST25Q']
	names(items29) <- c("Magazines", "Comic books", "Fiction",
	                   "Non-fiction books", "Newspapers")
	l29 <- likert(items29)
	l29
	lapply(items29, levels)

like this 

	u <- unique(scrounging$behav)
	levels.behav <- c("inactive", u[u!="inactive"])
	scrounging$behav <- factor(x=scrounging$behav, levels=levels.behav)
	u <- unique(scrounging$substr)
	levels.substr <- c("ground", u[u!="ground"])
	scrounging$substr <- factor(x=scrounging$substr, levels=levels.substr)
	likert::likert(scrounging)

yielding

	    Item inactive active foraging snoozing
	1  behav       36     20       20       24
	2 substr       20     28       30       22

but I doubt this is meaningful ... 

Best,
Rasmus

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-help/attachments/20210311/5e5a412e/attachment.sig>

  
  


  


      

    

    

      
      

        


  

        

      Richard M. Heiberger
    

    

      
      #
    

  


  

      
The likert() function in library(likert) is not the same as
the likert() function in library(HH).

The likert function in the likert package creates an object that needs to be plotted.
For the likert package you left out the line
        plot(likert::likert(scrounging))
This plot is not consistent with what your initial email said it was looking for.


Both Jim Lemon and I gave you plots that are consistent with your description.
You could probably get a similar plot from likert::likert(), but not by the lines you are using.
The likert::likert() function does not work with the table that HH::likert() and barplot() are using.
look at str(l29) to see what it needs

The likert package and the HH package cannot be loaded simultaneously.
Their use the same function names in incompatible ways.

How to get the ?likert::likert plot using HH:

HH::likert(t(sapply(items29, table)), as.percent=TRUE, positive.order=TRUE)

Here is a complete example using Jim Lemon's data.



## NO library() statements!
## The likert and HH packages have incompatible usage
## of the same function names.

## this is the example from ?likert::likert
data(pisaitems, package="likert")
items29 <- pisaitems[,substr(names(pisaitems), 1,5) == 'ST25Q']
names(items29) <- c("Magazines", "Comic books", "Fiction",
                    "Non-fiction books", "Newspapers")
## plot using likert package
l29 <- likert::likert(items29)
likert:::plot.likert(l29)


## plot using HH package
HH::likert(t(sapply(items29, table)), as.percent=TRUE,
           positive.order=TRUE, main="HH::likert")

HH::likert(t(sapply(items29, table)), as.percent=TRUE,
           positive.order=TRUE, main="HH::likert",
           auto.key=list(columns=2),
           scales=list(y=list(tck=c(0,2)))) ## prettier



## Jim Lemon's example
scrounging <- data.frame(
  behav=sample(c("inactive","active","foraging","snoozing"),50,TRUE),
  substr=sample(c("tree","ground","vine","air"),50,TRUE))

scroungeTable <- t(table(scrounging))
scroungeTable
HH::likert(scroungeTable)

## I don't think Jim Lemon's example can be used directly in the likert package.
## look at
str(l29)
## and notice that it must include both raw data and the table
l29$results
head(l29$items)


## For any further discussion on this list please include the output from
dput(head(yourRealData))
## in the body of the email.

## Jim's example is based on base graphics barplot
## The HH likert() function is based on lattice barchart.
## the likert package's likert:::plot.likert() function is based on ggplot.

        

      
      
    

  
  


  


      

    

    

      
      

        
      

    

    

      
      

        


  

        

      Areti Panopoulou
    

    

      
      #
    

  


  

      
Thank you so much everyone for your time and help! I really appreciate it.

Kind Regards,
Areti

        
On Thu, Mar 11, 2021 at 10:34 PM Rasmus Liland <jral at posteo.no> wrote: