We will be extracting tweets with the substring ‘Barclays’ from Twitter with the help of the ‘TwitteR’ package and use these tweets to perform Sentiment Analysis on it to find out the sentiments of the people on Barclays.

We will be making use of the NLP(Natural Language Processing) package along with the tm(Text Mining) package.

The twitter application for the Sentiment Analysis has already been made prior to writing this code.

library("tm")
## Loading required package: NLP
library("NLP")
library("Rstem")
library("twitteR")
library("sentiment")
library("plyr")
## 
## Attaching package: 'plyr'
## 
## The following object is masked from 'package:twitteR':
## 
##     id
library("ggplot2")
## Warning: package 'ggplot2' was built under R version 3.2.3
## 
## Attaching package: 'ggplot2'
## 
## The following object is masked from 'package:NLP':
## 
##     annotate
library("wordcloud")
## Loading required package: RColorBrewer
library("RColorBrewer")
library("ROAuth")
library("httk")
library("base64enc")
library("SnowballC")
## 
## Attaching package: 'SnowballC'
## 
## The following objects are masked from 'package:Rstem':
## 
##     getStemLanguages, wordStem
download.file(url="http://curl.haxx.se/ca/cacert.pem",destfile="cacert.pem")

#create an object "cred" that will save the authenticated object that we can use for later sessions
cred <- OAuthFactory$new(consumerKey='type ur consumer key',
                         consumerSecret='type ur consumer secret key',
                         requestURL='https://api.twitter.com/oauth/request_token',
                         accessURL='https://api.twitter.com/oauth/access_token',
                         authURL='https://api.twitter.com/oauth/authorize')

# Executing the next step generates an output --  To enable the connection, please direct your web browser to: <hyperlink  . Note:  You only need to do this part once

save(cred, file="twitter authentication.Rdata")

load("twitter authentication.Rdata")

customer_key <- 'ltpJUCVPf0dF5awEeMuICCubp'
customer_secret <- 'GcKbtzXhfGjIje3fL7DUY38warfkZ231D3cr8KKKXiN9t3iOAi'
access_token <- '4027888634-igfeIOAYSVExtwPoLo7pbA2V9oJ1pOih8W1ey1P'
access_secret <- '8McctpjrPBuVgHLuvAMrcKFLAkur42HK0YquqTDsFlwxl'

setup_twitter_oauth(customer_key,customer_secret,access_token,access_secret)
## [1] "Using direct authentication"
bjp_tweets = searchTwitter("Barclays", n=2000, lang="en")

# fetch the text of these tweets
bjp_txt = sapply(bjp_tweets, function(x) x$getText())

# Now we will prepare the above text for sentiment analysis

# First we will remove retweet entities from the stored tweets (text)
bjp_txt <- iconv(bjp_txt, to = "utf-8", sub="")

bjp_txt = gsub("(RT|via)((?:\\b\\W*@\\w+)+)", "", bjp_txt)

# Then remove all "@people"
bjp_txt = gsub("@\\w+", "", bjp_txt)

# Then remove all the punctuation
bjp_txt = gsub("[[:punct:]]", "", bjp_txt)

# Then remove numbers, we need only text for analytics
bjp_txt = gsub("[[:digit:]]", "", bjp_txt)

# the remove html links, which are not required for sentiment analysis
bjp_txt = gsub("http\\w+", "", bjp_txt)

# finally, we remove unnecessary spaces (white spaces, tabs etc)
bjp_txt = gsub("[ \t]{2,}", "", bjp_txt)
bjp_txt = gsub("^\\s+|\\s+$", "", bjp_txt)

# if anything else, you feel, should be removed, you can. For example "slang words" etc using the above function and methods.

# Since there can be some words in lower case and some in upper, we will try to eredicate this non-uniform pattern by making all the words in lower case. This makes uniform pattern.

# Let us first define a function which can handle "tolower error handling", if arises any, during converting all words in lower case.
catch.error = function(x)
{
        # let us create a missing value for test purpose
        y = NA
        # try to catch that error (NA) we just created
        catch_error = tryCatch(tolower(x), error=function(e) e)
        # if not an error
        if (!inherits(catch_error, "error"))
                y = tolower(x)
        # check result if error exists, otherwise the function works fine.
        return(y)
}

# Now we will transform all the words in lower case using catch.error function we just created above and with sapply function
bjp_txt = sapply(bjp_txt, catch.error)

# Also we will remove NAs, if any exists, from bjp_txt (the collected and refined text in analysis)
bjp_txt = bjp_txt[!is.na(bjp_txt)]

# also remove names (column headings) from the text, as we do not want them in the sentiment analysis
names(bjp_txt) = NULL

# Now the text is fully prepared (or at least for this tutorial) and we are good to go to perform Sentiment Analysis using this text

# As a first step in this stage, let us first classify emotions
# In this tutorial we will be using Baye's algorithm to classify emotion categories
# for more please see help on classify_emotion (?classify_emotion) under sentiment package
bjp_class_emo = classify_emotion(bjp_txt, algorithm="bayes", prior=1.0)
# the above function returns an of bject of class data.frame with seven columns (anger, disgust, fear, joy, sadness, surprise, best_fit) and one row for each document:

# we will fetch emotion category best_fit for our analysis purposes, visitors to this tutorials are encouraged to play around with other classifications as well.
emotion = bjp_class_emo[,7]

# Replace NA's (if any, generated during classification process) by word "unknown"
# There are chances that classification process generates NA's. This is because, sentiment package uses an in-built dataset "emotions", which containing approximately 1500 words classified into six emotion categories: anger, disgust, fear, joy, sadness, and surprise
# If any words outside this dataset are given, the process will term the words as NA's
emotion[is.na(emotion)] = "unknown"

# Similar to above, we will classify polarity in the text
# This process will classify the text data into four categories (The absolute log likelihood of the document expressing a positive sentiment, neg. The absolute log likelihood of the document expressing a negative sentiment, pos/neg. The ratio of absolute log likelihoods between positive and negative sentiment scores where a score of 1 indicates a neutral sentiment, less than 1 indicates a negative sentiment, and greater than 1 indicates a positive sentiment; AND best_fit. The most likely sentiment category (e.g. positive, negative, neutral) for the given text)

bjp_class_pol = classify_polarity(bjp_txt, algorithm="bayes")

# we will fetch polarity category best_fit for our analysis purposes, and as usual, visitors to this tutorials are encouraged to play around with other classifications as well
polarity = bjp_class_pol[,4]

# Let us now create a data frame with the above results obtained and rearrange data for plotting purposes
# creating data frame using emotion category and polarity results earlier obtained
sentiment_dataframe = data.frame(text=bjp_txt, emotion=emotion, polarity=polarity, stringsAsFactors=FALSE)

# rearrange data inside the frame by sorting it
sentiment_dataframe = within(sentiment_dataframe, emotion <- factor(emotion, levels=names(sort(table(emotion), decreasing=TRUE))))

Now that we have extracted all the tweets from Twitter and classified them with respect of their polarities, we will like to do some Data Visualization to get the insights on the sentiments of people on Barclays.

ggplot(sentiment_dataframe, aes(x=emotion)) + geom_bar(aes(y=..count.., fill=emotion)) +
        scale_fill_brewer(palette="Dark2") +
        ggtitle('Sentiment Analysis of Tweets on Twitter about Barclays') +
        theme(legend.position='right') + ylab('Number of Tweets') + xlab('Emotion Categories')

dev.off()
## null device 
##           1
ggplot(sentiment_dataframe, aes(x=polarity)) + geom_bar(aes(y=..count.., fill=polarity)) + scale_fill_brewer(palette="Set1") + ggtitle('Sentiment Analysis of Tweets on Twitter about Barclays') + theme(legend.position='right') + ylab('Number of Tweets') + xlab('Polarity Categories')
dev.off()
## null device 
##           1
swb_emos = levels(factor(sentiment_dataframe$emotion))
n_swb_emos = length(swb_emos)
swb.emo.docs = rep("", n_swb_emos)
for (i in 1:n_swb_emos)
{
        tmp = bjp_txt[emotion == swb_emos[i]]
        swb.emo.docs[i] = paste(tmp, collapse=" ")
}
swb.emo.docs = removeWords(swb.emo.docs, stopwords("english"))
swb.corpus = Corpus(VectorSource(swb.emo.docs))
swb.tdm = TermDocumentMatrix(swb.corpus)
swb.tdm = as.matrix(swb.tdm)
colnames(swb.tdm) = swb_emos

comparison.cloud(swb.tdm, colors = brewer.pal(8, "Dark2"),scale = c(1,.5), random.order = FALSE, title.size = 0.5)
dev.off()
## null device 
##           1