elza_soares.r

elza<- rtweet::search_tweets("elza", n=18000, include_rts = FALSE )

tidy_elza<- 
  elza %>%
  unnest_tokens(word, text, token = "tweets", strip_punct = TRUE, strip_url = TRUE) %>%
  group_by(word) %>%
  filter(n() > 5) %>%
  ungroup() %>%
  select(word)

library(stopwords) 

library(tibble)

stopword <- as_tibble(stopwords::stopwords("pt")) 
stopword <- rename(stopword, word=value)

newstopwords_elza <- tibble(word = c("é", "t","n","q","vc","h", "pra","via","ñ","vai"))

stop_words_users_elza<- 
  tidy_twitter %>%
  filter(str_detect(word, "@")) %>%
  select(word)

stop_words_users_elza_sem_arroba<- 
  tidy_twitter %>%
  filter(str_detect(word, "@")) %>%
  mutate(word=str_sub(word,2,200)) %>%
  select(word)


stopwords_full_elza <- bind_rows (stopword, stop_words_users_elza,newstopwords_elza, stop_words_users_elza_sem_arroba)

tb_elza <- anti_join(tidy_elza, stopwords_full_elza, by = 'word')

word_count_elza <- count(tb_elza, word, sort = TRUE)


##Nuvem de palavras
library(wordcloud)
tb_elza %>%
  filter(!word %in% c("elza","soares")) %>%
  count(word)%>%
  with(wordcloud(word,n,max.words = 150, colors=brewer.pal(6,"Dark2"),random.order=FALSE))