Create wordcloud from a data frame in R

问题:

I made a sample data frame. I try to make a wordcloud from the Projects column.

Hours<-c(2,3,4,2,1,1,3)
Project<-c("a","b","b","a","c","c","c")
Period<-c("2014-11-22","2014-11-23","2014-11-24","2014-11-22", "2014-11-23", "2014-11-23", "2014-11-24")
cd=data.frame(Project,Hours,Period)

Here is my code:

cd$Project<-as.character(cd$Project)
wordcloud(cd$Project,min.freq=1)

but I get the following error:

Error in strwidth(words[i], cex = size[i], ...) : invalid 'cex' value
In addition: Warning messages:
1: In max(freq) : no non-missing arguments to max; returning -Inf
2: In max(freq) : no non-missing arguments to max; returning -Inf

What am I doing wrong?

回答1:

I think you are missing the freq argument. You want to create a column indicating how often each project happened. I, therefore, transformed your data using count in the dplyr package.

library(dplyr)
library(wordcloud)

cd <- data.frame(Hours = c(2,3,4,2,1,1,3),
                 Project = c("a","b","b","a","c","c","c"),             
                 Period = c("2014-11-22","2014-11-23","2014-11-24",
                            "2014-11-22", "2014-11-23", "2014-11-23",
                            "2014-11-24"),
                 stringsAsFactors = FALSE)

cd2 <- count(cd, Project)

#  Project n
#1       a 2
#2       b 2
#3       c 3

wordcloud(words = cd2$Project, freq = cd2$n, min.freq = 1)

回答2:

If you specify a character column, then the function creates a corpus and a document term matrix for you behind the scenes. The problem is that the default behavior for the TermDocumentMatrix function from the tm pacakge is that is only tracks words that are longer than three characters (also, it removes "stop words" so values like "a" would be removed). So if you changed your sample to

Project<-c("aaa","bbb","bbb","aaa","ccc","ccc","ccc")

it would work just fine. It appears there are no ways to change the control options sent to TermDocumentMatrix. If you want to calculate the frequencies yourself in the same way that the default wordcloud function does, you can do

corpus <- Corpus(VectorSource(cd$Project))
corpus <- tm_map(corpus, removePunctuation)
# corpus <- tm_map(corpus, function(x) removeWords(x, stopwords()))
tdm <-TermDocumentMatrix(corpus, control=list(wordLengths=c(1,Inf)))
freq <- slam::row_sums(tdm)
words <- names(freq)    

wordcloud(words, freq, min.freq=1)

however, for simple cases, you could just count the frequencies with table()

tbl <- table(cd$Project)
wordcloud(names(tbl), tbl, min.freq=1)