How to Perform Sentiment Analysis of Twitter Data using R Programming

Sentiment Analysis of Twitter Data using R Programming – To perform Sentiment Analysis in R we need “sentiment” R package well developed by Timothy P. Jurka. The Package comes with two important functions namely: “classify_emotion” and “classify_polarity”. In this post I will show How to use both the functions in brief:

Read Also:

First extract the tweet using the desired keyword. For demonstration process I am fetching 2000 tweets for both the keywords. So a total of 4000 tweets to analyze using sentiment package in R.

Read Fetching Twitter Data using R and Cleaning of Twitter Data.

In this tutorial I will use both “classify_emotion” and “classify_polarity” function to perform sentiment analysis on Twitter Data collected. See the code of the sentiment analysis function below:

sentiment_twitter <- function(searchterm,i)
{
  library(twitteR)
  library(wordcloud)
  library(tm)
  library(RColorBrewer)
  library(sentiment)
  library(plyr)
  library(ggplot2)
  consumer_key <- 'Your-Key'
  consumer_secret <- 'Your Secret'
  access_token<-'Your Access Token'
  access_secret <- 'Your Access Secret'
  setup_twitter_oauth(consumer_key, consumer_secret, access_token, access_secret)
  tweet_to_score <- searchTwitter(searchterm, n=i, lang="en", resultType="recent")
  tweet_text <- sapply(tweet_to_score, function(x) x$getText())
  tweet_clean <- clean_text(tweet_text)
  emotion_class <- classify_emotion(tweet_clean, algorithm = "bayes", prior = 1)
  emotion <- emotion_class[,7]
  emotion[is.na(emotion)] = "unknown"
  polarity_class <- classify_polarity(tweet_clean, algorithm = "bayes")
  polarity = polarity_class[,4]
  sent_df = data.frame(text=tweet_clean, emotion=emotion,polarity=polarity, stringsAsFactors=FALSE)
  sent_df = within(sent_df,emotion <- factor(emotion, levels=names(sort(table(emotion), decreasing=TRUE))))
  head(sent_df, n=5)
  ggplot(sent_df, aes(x=emotion)) +
    geom_bar(aes(y=..count.., fill=emotion)) +
    scale_fill_brewer(palette="Dark2") +
    labs(x="emotion categories", y="number of tweets") +
    ggtitle("Sentiment Analysis of Twitter") +
    theme(plot.title = element_text(size=12, face="bold"))
  }

Code of clean_text function

clean_text = function(x)
{
  x = gsub("rt", "", x) # remove rt
  x = gsub("RT", "", x) # remove RT
  x = gsub("@\\w+", "", x) # remove at
  x = gsub("[[:punct:]]", "", x) # remove punctuation
  x = gsub("[[:digit:]]", "", x) # remove numbers
  x = gsub("http\\w+", "", x)  # remove links http
  x = gsub("[ |\t]{2,}", "", x) # remove tabs
  x = gsub("^ ", "", x)  # remove blank spaces at the beginning
  x = gsub(" $", "", x) # remove blank spaces at the end
  try.error = function(z)
  {
    y = NA
    try_error = tryCatch(tolower(z), error=function(e) e)
    if (!inherits(try_error, "error"))
      y = tolower(z)
    return(y)
  }
  x = sapply(x, try.error)
  return(x)
}

Copy Paste the above code in a R Script File and call the function sentiment_twitter(“Search Term”, No. of tweets)

For example I am calling the function for search term: IPL2016