#This R script creates wordcloud of CRAN package titles since 2020.

wordcloud_example_CRAN_package_titles.r

#This R script creates wordcloud of CRAN package titles since 2020.
library(rvest)
library(ggplot2)
library(dplyr)
library(tokenizers)
library(SnowballC)
library(stopwords)
library(ggwordcloud)

URL<- "https://cran.r-project.org/web/packages/available_packages_by_date.html"
#Scrape table data from CRAN website
getthis<- xml2::read_html(URL) 
cran_df<-rvest::html_table(getthis) %>% data.frame(.)
#filter to packages published since 2020
cran_df_since2020<- cran_df[as.Date(cran_df$Date)>"2020-01-01",]

#concatenate title text, and removed some of the symbols in the text
CRAN_text<- cran_df_since2020 %>%
  summarize(test = toString(Title)) %>%
  ungroup() %>%
  gsub("\n", " ",.) %>%
  gsub("\\(|\\)|, ", " ",.)

#Tokenize text by wordstems, filter stopwords.
CRAN_text_token<-tokenize_word_stems(CRAN_text, stopwords = stopwords::stopwords("en"))

#Create a frequency table
CRAN_text_table<- table(CRAN_text_token) %>% data.frame(.) %>%
  arrange(desc(Freq))

#Limit frequency table to top 200 rows
CRAN_text_top200<-CRAN_text_table[1:200,]

#today's date (for labeling chart)
today_is<-Sys.Date()

#plot wordcloud.
p<-ggplot(CRAN_text_top200, aes(label = CRAN_text_token, size = Freq)) +
  geom_text_wordcloud() +
  scale_size_area(max_size = 36) +
  theme_minimal()+
  ggtitle("CRAN package vocabulary ")+
  labs(subtitle = paste0("Top 200 words in package titles, sized by frequency as of ", today_is))

p<-p+ theme(plot.title = element_text(size=32, color="#6699cc"),
            plot.subtitle = element_text(size=16, color="#6699cc"))

#Save to file
png(file="cran_wordcloud.png",
    width=1200, height=700)
p
dev.off()