[R] Different stack barplots - same color legends
Richard M. Heiberger
rmh @end|ng |rom temp|e@edu
Wed Oct 31 20:28:08 CET 2018
What you sent looks like a set of column labels, not the actual numeric data.
You might want to convert them to factors where you control the order
of the levels.
> Factor.wrong <- factor(c("mm", "cm", "m", "km"))
> levels(Factor.wrong) ## alphabetical order, not meaning order
[1] "cm" "km" "m" "mm"
>
> Factor.right <- factor(c("mm", "cm", "m", "km"),
+ levels=c("mm", "cm", "m", "km"))
> levels(Factor.right) ## meaning order
[1] "mm" "cm" "m" "km"
Or you might want to construct a matrix of counts of your data and plot that.
Rich
On Wed, Oct 31, 2018 at 1:53 PM, P. Roberto Bakker
<robertobakker using gmail.com> wrote:
> This is part of the output text
>
> "de situatie in hoge mate van toepassing is voor u of uw supervisorengroep",
> "de situatie in hoge mate van toepassing is voor u of uw supervisorengroep",
> "de situatie enigszins van toepassing is voor u of uw supervisorengroep",
> "de situatie in hoge mate van toepassing is voor u of uw supervisorengroep"
> ), `Ik waardeer de inbreng van de aios in de afdelingsvergadering` = c("de
> situatie in hoge mate van toepassing is voor u of uw supervisorengroep",
> "de situatie in hoge mate van toepassing is voor u of uw supervisorengroep",
> "de situatie in zeer hoge mate van toepassing is voor u of uw
> supervisorengroep",
> "de situatie in zeer hoge mate van toepassing is voor u of uw
> supervisorengroep",
> "de situatie enigszins van toepassing is voor u of uw supervisorengroep",
> "de situatie in zeer hoge mate van toepassing is voor u of uw
> supervisorengroep",
> "de situatie in zeer hoge mate van toepassing is voor u of uw
> supervisorengroep",
> "de situatie in zeer hoge mate van toepassing is voor u of uw
> supervisorengroep",
> "de situatie enigszins van toepassing is voor u of uw supervisorengroep",
> "de situatie in hoge mate van toepassing is voor u of uw supervisorengroep",
> "de situatie in hoge mate van toepassing is voor u of uw supervisorengroep",
> "de situatie in hoge mate van toepassing is voor u of uw supervisorengroep",
>
> Op wo 31 okt. 2018 om 16:24 schreef Richard M. Heiberger <rmh using temple.edu>:
>>
>> part is fine. just be sure that the small part causes the problem.
>> I will need that to investigate what is happening.
>>
>>
>> On Wed, Oct 31, 2018 at 11:15 AM, P. Roberto Bakker
>> <robertobakker using gmail.com> wrote:
>> > It is a very long result text. I can send it to you, or is part of it
>> > ok?[
>> >
>> > Op wo 31 okt. 2018 om 14:27 schreef Richard M. Heiberger
>> > <rmh using temple.edu>:
>> >>
>> >> Please send me the
>> >> dput(teamq)
>> >>
>> >>
>> >> On Wed, Oct 31, 2018 at 03:51 P. Roberto Bakker
>> >> <robertobakker using gmail.com>
>> >> wrote:
>> >>>
>> >>> Thank you for you information. Package 'HH' is interesting.
>> >>>
>> >>> Now I find another problem when using 'likert(teamq)'
>> >>> I get an error message:
>> >>> > likert(teamq)
>> >>> Error in dimnames(x) <- `*vtmp*` :
>> >>> length of 'dimnames' [2] not equal to array extent
>> >>>
>> >>> I checked:
>> >>> > dim(teamq)
>> >>> [1] 4 2
>> >>> > ncol(teamq)
>> >>> [1] 2
>> >>> So it should be good.
>> >>>
>> >>> I used 'make.names' , in case the spaces in the variable names would
>> >>> be a
>> >>> problem.
>> >>> Same error.
>> >>>
>> >>> What could I do?
>> >>>
>> >>> Best and thank you in advance.
>> >>> Roberto
>> >>>
>> >>>
>> >>> Op ma 22 okt. 2018 om 20:10 schreef Richard M. Heiberger
>> >>> <rmh using temple.edu>:
>> >>>>
>> >>>> Try the likert function in
>> >>>> install.packages("HH) ## if necessary
>> >>>> library(HH)
>> >>>>
>> >>>> Then using David Carlson's example teamq
>> >>>> likert(teamq)
>> >>>>
>> >>>> Your example in the 1:30PM (Eastern Daylight Time) doesn't work.
>> >>>> Error in revalue(teamq, c(`de situatie in zeer geringe mate van
>> >>>> toepassing is\nvoor u of uw supervisorengroep` = "1", :
>> >>>> x is not a factor or a character vector.
>> >>>>
>> >>>> There are many examples in
>> >>>> ?likert
>> >>>>
>> >>>> Rich
>> >>>>
>> >>>>
>> >>>> On Mon, Oct 22, 2018 at 1:30 PM, P. Roberto Bakker
>> >>>> <robertobakker using gmail.com> wrote:
>> >>>> > Dear David,
>> >>>> >
>> >>>> > Thank you for you quite response.
>> >>>> > My apologies for not giving some sample data - this is due to AVG.
>> >>>> > *But this minisample should not be a problem (all in Dutch)*:
>> >>>> > teamq
>> >>>> > # A tibble: 4 x 2
>> >>>> > `Ik volg bijscholing om mijn opleiders-kwaliteiten op peil te
>> >>>> > houden`
>> >>>> >
>> >>>> > <chr>
>> >>>> >
>> >>>> > 1 de situatie in hoge mate van toepassing is voor u of uw
>> >>>> > supervisorengroep
>> >>>> > 2 de situatie in zeer hoge mate van toepassing is voor u of uw
>> >>>> > supervisorengroep
>> >>>> > 3 de situatie in zeer hoge mate van toepassing is voor u of uw
>> >>>> > supervisorengroep
>> >>>> > 4 de situatie in geringe mate van toepassing is voor u of uw
>> >>>> > supervisorengroep
>> >>>> > `Ik weet precies wat de ‘modernisering van de opleiding’ inhoudt`
>> >>>> >
>> >>>> > <chr>
>> >>>> >
>> >>>> > 1 de situatie in hoge mate van toepassing is voor u of uw
>> >>>> > supervisorengroep
>> >>>> > 2 de situatie in zeer hoge mate van toepassing is voor u of uw
>> >>>> > supervisorengroep
>> >>>> > 3 de situatie in zeer hoge mate van toepassing is voor u of uw
>> >>>> > supervisorengroep
>> >>>> > 4 de situatie in geringe mate van toepassing is voor u of uw
>> >>>> > supervisorengroep
>> >>>> >
>> >>>> > As you see the likert items are in words, and I should change them
>> >>>> > in
>> >>>> > nummeric - Am I correct?
>> >>>> >
>> >>>> > *To do this, I tried (see further below):*
>> >>>> > plyr rename() ; I receive the message it should be a factor or
>> >>>> > character
>> >>>> > dplyr recode() ; same message
>> >>>> > mapvalues() ; it should be atomic, so I used as.atomic(teamq) but
>> >>>> > then
>> >>>> > I
>> >>>> > receive the nummers a strings.
>> >>>> >
>> >>>> > *The syntaxes*
>> >>>> > require(plyr)
>> >>>> > example2 <- revalue(teamq,
>> >>>> > c("de situatie in zeer geringe mate van
>> >>>> > toepassing
>> >>>> > is
>> >>>> > voor u of uw supervisorengroep"= "1",
>> >>>> > "de situatie in geringe mate van toepassing
>> >>>> > is
>> >>>> > voor
>> >>>> > u of uw supervisorengroep"= "2",
>> >>>> > "de situatie enigszins van toepassing is
>> >>>> > voor
>> >>>> > u of
>> >>>> > uw supervisorengroep"= "3",
>> >>>> > "de situatie in hoge mate van toepassing is
>> >>>> > voor u
>> >>>> > of uw supervisorengroep"= "4",
>> >>>> > "de situatie in zeer hoge mate van
>> >>>> > toepassing
>> >>>> > is
>> >>>> > voor u of uw supervisorengroep"= "5"))
>> >>>> >
>> >>>> > require(dplyr)
>> >>>> > example2 <- recode(teamq,
>> >>>> > c("de situatie in zeer geringe mate van
>> >>>> > toepassing
>> >>>> > is
>> >>>> > voor u of uw supervisorengroep"= "1",
>> >>>> > "de situatie in geringe mate van toepassing
>> >>>> > is
>> >>>> > voor u
>> >>>> > of uw supervisorengroep"= "2",
>> >>>> > "de situatie enigszins van toepassing is voor
>> >>>> > u
>> >>>> > of uw
>> >>>> > supervisorengroep"= "3",
>> >>>> > "de situatie in hoge mate van toepassing is
>> >>>> > voor
>> >>>> > u of
>> >>>> > uw supervisorengroep"= "4",
>> >>>> > "de situatie in zeer hoge mate van toepassing
>> >>>> > is
>> >>>> > voor
>> >>>> > u of uw supervisorengroep"= "5"))
>> >>>> >
>> >>>> > mapvalues(as.matrix(teamq), from = c("de situatie in zeer geringe
>> >>>> > mate
>> >>>> > van
>> >>>> > toepassing is voor u of uw supervisorengroep",
>> >>>> > "de situatie in geringe mate van toepassing
>> >>>> > is
>> >>>> > voor
>> >>>> > u of uw supervisorengroep",
>> >>>> > "de situatie enigszins van toepassing is
>> >>>> > voor
>> >>>> > u of
>> >>>> > uw supervisorengroep",
>> >>>> > "de situatie in hoge mate van toepassing is
>> >>>> > voor u
>> >>>> > of uw supervisorengroep",
>> >>>> > "de situatie in zeer hoge mate van
>> >>>> > toepassing
>> >>>> > is
>> >>>> > voor u of uw supervisorengroep"),
>> >>>> > to = c(1,2,3,4,5))
>> >>>> >
>> >>>> > What should I do?
>> >>>> > Thank you in advance, Roberto
>> >>>> >
>> >>>> > Op ma 22 okt. 2018 om 17:13 schreef David L Carlson
>> >>>> > <dcarlson using tamu.edu>:
>> >>>> >
>> >>>> >> Your example is not reproducible since you did not give us some
>> >>>> >> sample
>> >>>> >> data. I suspect that your data frame consists of columns that
>> >>>> >> represent
>> >>>> >> questions and rows that represent individuals who answered the
>> >>>> >> questions.
>> >>>> >> First create a simple example:
>> >>>> >>
>> >>>> >> set.seed(42)
>> >>>> >> teamq <- data.frame(V1=sample(c(1, 2, 4, 5), 25, replace = TRUE),
>> >>>> >> V2=sample(c(1, 2, 3, 4, 5), 25, replace=TRUE),
>> >>>> >> V3=sample(c(2, 3, 4, 5), 25, replace=TRUE))
>> >>>> >>
>> >>>> >> Notice that this data frame ONLY contains questions (and only 3
>> >>>> >> questions). Here are 2 ways to get what you want. The first one
>> >>>> >> stacks the
>> >>>> >> data:
>> >>>> >>
>> >>>> >> teamq.stack <- stack(teamq)
>> >>>> >> str(teamq.stack)
>> >>>> >> counts <- table(teamq.stack)
>> >>>> >> str(counts)
>> >>>> >>
>> >>>> >> The second one converts each column to a factor with levels 1 - 5:
>> >>>> >>
>> >>>> >> teamq2 <- data.frame(lapply(teamq, factor, levels=1:5))
>> >>>> >> str(teamq2)
>> >>>> >> counts <- sapply(teamq2, table)
>> >>>> >> str(counts)
>> >>>> >>
>> >>>> >> Now make the plots:
>> >>>> >>
>> >>>> >> cols <- c("yellow","sandybrown","orange",
>> >>>> >> "darkolivegreen","green")
>> >>>> >> barplot(counts[, 1], horiz=TRUE, col=cols, legend=TRUE)
>> >>>> >> barplot(counts[, 2], horiz=TRUE, col=cols, legend=TRUE)
>> >>>> >> barplot(counts[, 3], horiz=TRUE, col=cols, legend=TRUE)
>> >>>> >>
>> >>>> >> You will need to adjust the xlim= argument so that the legend does
>> >>>> >> not
>> >>>> >> print on top of the bars.
>> >>>> >>
>> >>>> >> ----------------------------------------
>> >>>> >> David L Carlson
>> >>>> >> Department of Anthropology
>> >>>> >> Texas A&M University
>> >>>> >> College Station, TX 77843-4352
>> >>>> >>
>> >>>> >>
>> >>>> >> -----Original Message-----
>> >>>> >> From: R-help <r-help-bounces using r-project.org> On Behalf Of P.
>> >>>> >> Roberto
>> >>>> >> Bakker
>> >>>> >> Sent: Monday, October 22, 2018 9:04 AM
>> >>>> >> To: R mailing list <r-help using r-project.org>
>> >>>> >> Subject: [R] Different stack barplots - same color legends
>> >>>> >>
>> >>>> >> Hi,
>> >>>> >>
>> >>>> >> I want to make barplots from different questions (columns) in one
>> >>>> >> data.frame.
>> >>>> >> Each question has the same 5 likert items.
>> >>>> >> Now the problem: in some questions all items are answered; in
>> >>>> >> other
>> >>>> >> less.
>> >>>> >> From the syntax below I get nice stack barplots - *but the legend
>> >>>> >> colors do
>> >>>> >> not* refer to the same likert-item, which I understand - the
>> >>>> >> colors
>> >>>> >> go in
>> >>>> >> sequence along the table.
>> >>>> >> Question: how can I write a syntax that each likert-item has the
>> >>>> >> same
>> >>>> >> legend color?
>> >>>> >> Thank you in advance,
>> >>>> >>
>> >>>> >> Roberto
>> >>>> >>
>> >>>> >> SYNTAX:
>> >>>> >> counts19 <- table(teamq[,19])
>> >>>> >> counts20 <- table(teamq[,20])
>> >>>> >> barplot(as.matrix(counts19), horiz = T,
>> >>>> >> col=c("yellow","sandybrown","orange",
>> >>>> >> "darkolivegreen","green"),
>> >>>> >> legend=T)
>> >>>> >> barplot(as.matrix(counts20), horiz = T,
>> >>>> >> col=c("yellow","sandybrown","orange",
>> >>>> >> "darkolivegreen","green"),
>> >>>> >> legend=T)
>> >>>> >>
>> >>>> >> [[alternative HTML version deleted]]
>> >>>> >>
>> >>>> >> ______________________________________________
>> >>>> >> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> >>>> >> https://stat.ethz.ch/mailman/listinfo/r-help
>> >>>> >> PLEASE do read the posting guide
>> >>>> >> http://www.R-project.org/posting-guide.html
>> >>>> >> and provide commented, minimal, self-contained, reproducible code.
>> >>>> >>
>> >>>> >
>> >>>> > [[alternative HTML version deleted]]
>> >>>> >
>> >>>> > ______________________________________________
>> >>>> > R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> >>>> > https://stat.ethz.ch/mailman/listinfo/r-help
>> >>>> > PLEASE do read the posting guide
>> >>>> > http://www.R-project.org/posting-guide.html
>> >>>> > and provide commented, minimal, self-contained, reproducible code.
More information about the R-help
mailing list