#calculate the sentiment score
sentiment_tweets<-sentiment(tweets.df$text)


#adding sentiment score to df
tweets.df<-tweets.df%>%mutate(sentiment=sentiment_tweets$sentiment)
View(tweets.df)


#get the positive sentiment score
positive_tweets<-head(arrange(unique(tweets.df[,c(1,17)]),desc(sentiment)),30)
positive_tweets
write.table(positive_tweets$text,file='E:/Program Files/RStudio/New folder/pos_tweets.txt',sep='n')


#get the negative sentiment score
negative_tweets<-head(arrange(unique(tweets.df[,c(1,17)]),sentiment),30)
negative_tweets
write.table(positive_tweets$text,file='E:/Program Files/RStudio/New folder/neg_tweets.txt',sep='n')


#get the document
mycorpus2<-Corpus(DirSource(directory='E:/Program Files/RStudio/New folder'))
summary(mycorpus2)


#clean the document
mycorpus2<-tm_map(mycorpus2,tolower)
mycorpus2<-tm_map(mycorpus2,removePunctuation)
mycorpus2<-tm_map(mycorpus2,removeWords,stopwords())
mycorpus2<-tm_map(mycorpus2,stripWhitespace)
mycorpus2<-tm_map(mycorpus2,stemDocument)



#convert corpus(class) into tdm
mycorpus2_tdm<-TermDocumentMatrix(mycorpus2)


#convert tdm into matrix
mycorpus2_matrix<-as.matrix(mycorpus2_tdm)

#name the columns
colnames(mycorpus2_matrix)<-c('negative','positive')
head(mycorpus2_matrix)


#graphing the key words
comparison.cloud(mycorpus2_matrix,max.words=100,random.order=F)