Text Mining Twitter

## Acceso a la Twitter API
# for this method you need the following objects from the
# "keys and access token" tab in your developers account

key <-"3omnRf2j6bp9AscKkf0VMlqsk"
secret <- "V22lCy4eWZNf7tp853Cjlxcpnpsot24EwnH0txxFilWx9EmloV"
secrettk <- "H0xTlxFuxZ2Qae9S2nKJYwWjl5CTvEcj2HHsCY5N1fWj1"
mytoken <- "301911069-Jfw1ZWsOr7scILP7Zjt7xTU5Dpb7cM0wdqEPb8V1"

#packages
library("twitteR")
library("httr")

# keep this order of arguments
setup_twitter_oauth(key, secret, mytoken, secrettk)

### Let’s start with the Twitter scraping!
library("twitteR")

# we are now scraping x tweets for Udemy
udemytweets = searchTwitter("#Udemy", n=100, lang="en")

library("tm")

udemylist <- sapply(udemytweets, function(x) x$getText())

# initiating a function

udemycorpus <- Corpus(VectorSource(udemylist))

## Transformation
# putting text to lower case
udemycorpus <- tm_map(udemycorpus, content_transformer(tolower))
# remove puntuación
udemycorpus <- tm_map(udemycorpus, removePunctuation)
# remove stopwords (meaningless words)
udemycorpus <- tm_map(udemycorpus, function(x)removeWords(x,stopwords()))

# keep a copy of corpus to use later as a dictionary for stem completion
#udemycorpuscopy <- udemycorpus

# stem words
#udemycorpus <- tm_map(udemycorpus, stemDocument)

# stem completion
#udemycorpus <- tm_map(udemycorpus, stemCompletion, dictionary = udemycorpuscopy)

# quitar espacios en blanco
udemycorpus <- tm_map(udemycorpus, stripWhitespace)

# to trasform to plain text which wordcloud can use
udemycorpus <- tm_map(udemycorpus, PlainTextDocument)

library("wordcloud")

palette <- brewer.pal(5,"Paired")
wordcloud(udemycorpus, min.freq=3, scale=c(4,1),
          random.color=F, max.word=45, random.order=F, colors=palette, rot.per=.5)