文本挖掘-R中事务活动的WordCloud

Question

我正试图从一些交易活动中生成一个wordcloud，以显示人们在哪里花的钱最多。交易活动如下所示：

Description       Amount
Albertson         20
Albertson         30
Albertson         35
CVS               10
CVS               40
Walmart           15
Walmart           44
...

我可以通过Description的频率轻松生成wordcloud。但我如何才能得到按每个类别的总和（数量）排序的wordcloud？谢谢

BTW这是我的代码

require(tm)
require(wordcloud)
require(RColorBrewer)
data_corpus <- Corpus(VectorSource(data))
data_corpus <- tm_map(data_corpus, content_transformer(tolower), mc.cores=1)
data_corpus <- tm_map(data_corpus, removePunctuation, mc.cores=1)
data_corpus <- tm_map(data_corpus, function(x)removeWords(x,stopwords()), mc.cores=1)
data_corpus <- tm_map(data_corpus, removeNumbers, mc.cores=1)
pal2 <- brewer.pal(8,"Dark2")
png("25-34.png", width=1280,height=800)
wordcloud(data_corpus, scale=c(6,.2),min.freq=50,max.words=Inf, random.order=FALSE, rot.per=.15, colors=pal2)
dev.off()

Answer 1

我将您的迷你表加载到一个名为data的数据帧中。然后运行以下代码：

require(wordcloud)
require(RColorBrewer)
library(dplyr)
# group by Description and sum the Amounts
data <- data %>% group_by(Description) %>% summarise(Amount = sum(Amount))
pal2 <- brewer.pal(8,"Dark2")
wordcloud(data$Description, freq = data$Amount, scale=c(6,.2),min.freq=50,max.words=Inf, random.order=FALSE, rot.per=.15, colors=pal2)

不需要tm软件包。只需在单词部分指定您的描述，在频率部分指定金额即可。