Sentiment Analysis using Sentimentr Package/Sentiment Analysis

if (!require("sentimentr"))install.packages('sentimentr')

## Loading required package: sentimentr

if (!require("dplyr"))install.packages('dplyr')

## Loading required package: dplyr

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(sentimentr)
library(dplyr)

Get sentiment of the sentence using function sentiment() to understand polarity of text(if it is a positive or negative)

text<- "I am enjoying my coffee. I think it's a perfect day today."

sentiment(text)

sentiment_by(text)

emotion(text)

Get aggregate of sentiment usign sentiment_by() function to understand the total polarity of text

stext<- "I am feeling terrible. Why does this happen to only me?"

sentiment(stext)

sentiment_by(stext)

emotion(stext)

Get emotion of the sentence using function emotion() to understand how the writer/author of sentence feels based on text

stext2<- "I am feeling terrible. Hope it gets better tomorrow."
sentiment(stext2)

sentiment_by(stext2)

emotion(stext2)

Detect profanity if text using function profanity()

ptext<- "The movie is just ridiculous!. I think it's shitty writing."
sentiment(ptext)

profanity(ptext)

Get Sentences from a given text as list using get_sentences() method

sent <- get_sentences(text)
sent

## [[1]]
## [1] "I am enjoying my coffee."          "I think it's a perfect day today."
## 
## attr(,"class")
## [1] "get_sentences"           "get_sentences_character"
## [3] "list"

How many kinds of lexicons does ‘sentimentr’ package have? How are they different from each other?

nrc
bing
afinn

$Difference in Lexicons$

Difference in Lexicons

Consider the ‘nrc’ package of ‘sentimentr’ package and display the count of the number of words of each kind of sentiment.

library(tidytext)
if(!require(textdata))install.packages('textdata')

## Loading required package: textdata

words_nrc<-get_sentiments("nrc")
words_nrc

words_nrc %>% group_by(sentiment)%>%
  count()

10.c Apply the ‘sentiment’function to ‘I am not very successful. He is successful’ and displays the result.

c= "I am not very successful. He is successful"
sentiment(c)

What’s the use of ‘extract_sentiment_terms’ function? Consider the following sentence and apply this function. ‘My life has become terrible since I met you and lost money. But I still have got a little hope left in me’.

extract_sentiment_terms

d <- "My life has become terrible since I met you and lost money. But I still have got a little hope left in me"

sentences_d <- get_sentences(d)
sentences_d

## [[1]]
## [1] "My life has become terrible since I met you and lost money."
## [2] "But I still have got a little hope left in me"              
## 
## attr(,"class")
## [1] "get_sentences"           "get_sentences_character"
## [3] "list"

pol_d <-extract_sentiment_terms(sentences_d)
pol_d

#pol$sentence
#pol$neutral
#pol$positive
#pol$negative

data.table::as.data.table(pol_d)

sentiment_counts_d<-attributes(pol_d)$counts
sentiment_counts_d

sentiment_elements_d<-attributes(pol_d)$elements
sentiment_elements_d

Build a word cloud after fetching the positive words using extract_sentiment_terms()

if(!require(data.table))install.packages('data.table')

## Loading required package: data.table

## 
## Attaching package: 'data.table'

## The following objects are masked from 'package:dplyr':
## 
##     between, first, last

library(wordcloud)

## Loading required package: RColorBrewer

library(data.table)

sentiment_counts_d[polarity > 0,]

par(mfrow = c(1, 3), mar = c(0, 0, 0, 0))
## Positive Words
with(
    sentiment_counts_d[polarity > 0,],
    wordcloud(words = words, freq = n, min.freq = 1,
          max.words = 200, random.order = FALSE, rot.per = 0.35,
          colors = brewer.pal(8, "Dark2"), scale = c(3, .75)
    )
)
mtext("Positive Words", side = 4, padj = 5)

## Negative Words
sentiment_counts_d[polarity < 0,]

with(
    sentiment_counts_d[polarity < 0,],
    wordcloud(words = words, freq = n, min.freq = 1,
          max.words = 200, random.order = FALSE, rot.per = 0.35,
          colors = brewer.pal(8, "Dark2"), scale = c(3, .75)
    )
)
mtext("Negative Words", side = 2, padj = 5)

## Neutral Words
sentiment_counts_d[polarity == 0,]

with(
    sentiment_counts_d[polarity == 0,],
    wordcloud(words = words, freq = n, min.freq = 1,
          max.words = 200, random.order = FALSE, rot.per = 0.35,
          colors = brewer.pal(8, "Dark2"), scale = c(3, .75)
    )
)
mtext("Neutral Words", side = 2, padj = 5)

Consider the following sentences and highlight the positive and negative sentences in different colors. Also, plot a graph displaying the variation of sentiments of the given sentences. ‘Rekha is an obedient daughter’,’I have lost hope in my life’, the fan has stopped moving’,’Riya looks beautiful’.

e<-"Rekha is an obedient daughter.I have lost hope in my life. the fan has stopped moving.Riya looks beautiful"
sentences_e <-get_sentences(e)
sentences_e

## [[1]]
## [1] "Rekha is an obedient daughter." "I have lost hope in my life."  
## [3] "the fan has stopped moving."    "Riya looks beautiful"          
## 
## attr(,"class")
## [1] "get_sentences"           "get_sentences_character"
## [3] "list"

Sentiment of 4 sentences are as follows

if(!require('dplyr'))install.packages('dplyr')
library(dplyr)
sentiment_e<-sentiment(sentences_e)
sentiment_e

sentiment_e_tbl<-data.table::as.data.table(sentiment_e)

Extracting sentiment terms for a given a list of sentences.

pol_e <-extract_sentiment_terms(sentences_e)
pol_e

sentiment_counts_e<-attributes(pol_e)$counts
sentiment_elements_e<-attributes(pol_e)$elements
pol_e_tbl<-data.table::as.data.table(sentiment_elements_e)
pol_e_tbl

Positive and Negative list of sentences filtered

positive_sen <-sentiment_e_tbl %>%filter(sentiment>0) 
negative_sen <-sentiment_e_tbl %>%filter(sentiment<0)

Highlighted Sentences Based on Sentiment

if(!require('ggplot2'))install.packages('ggplot2')

## Loading required package: ggplot2

library(ggplot2)

sentences_e_tbl <-data.table::as.data.table(sentences_e)
sentences_e_tbl %>%
  get_sentences()%>%
  sentiment_by()%>%
  highlight()

## Saved in /var/folders/q1/tpbkm_bj6dv12y3kmgdzq8280000gn/T//Rtmp5IoVyf/polarity.html

## Opening /var/folders/q1/tpbkm_bj6dv12y3kmgdzq8280000gn/T//Rtmp5IoVyf/polarity.html ...

highlight() function generates below image on a browser and I have saved that image for reference purpose only.

Highlighted Sentences Based on Sentiment

Plot the sentiment analysis

The plot method for the class sentiment uses syuzhet’s get_transformed_values combined with ggplot2 to make a reasonable, smoothed plot for the duration of the text based on percentage, allowing for comparison between plots of different texts. This plot gives the overall shape of the text’s sentiment.

sentiment_counts_e

(out <- with(
    sentences_e_tbl, 
    sentiment_by(
        get_sentences(sentences_e_tbl) 
    )
))

plot(uncombine(out))

sentences_e_tbl

library(ggplot2)
if(!require('tidytext'))install.packages('tidytext')
if(!require('tidyr'))install.packages('tidyr')

## Loading required package: tidyr

library(tidytext)
library(tidyr)
bigram <- sentences_e_tbl %>%
  unnest_tokens(bigram, V1, token = "ngrams", n = 2) %>%
  filter(!is.na(bigram))

trigram <- sentences_e_tbl %>%
  unnest_tokens(trigram, V1, token = "ngrams", n = 3) %>%
  filter(!is.na(trigram))

bigrams_separated <- bigram %>%
  separate(bigram, c("word1", "word2"), sep = " ")
bigrams_separated

AFINN <- get_sentiments("afinn")
bigram_w <- bigrams_separated %>%
  filter(word1 != "not") %>%
  inner_join(AFINN, by = c(word2 = "word")) %>%
  count(word2, value, sort = TRUE)

bigram_w %>%
  mutate(contribution = n * value) %>%
  arrange(desc(abs(contribution))) %>%
  head(20) %>%
  mutate(word2 = reorder(word2, contribution)) %>%
  ggplot(aes(n * value, word2, fill = n * value > 0)) +
  geom_col(show.legend = FALSE) +
  labs(x = "Sentiment value * number of occurrences",
       y = "Words preceded by \"not\"")

bigrams_united <- bigrams_separated %>%
  unite(bigram, word1, word2, sep = " ")
bigrams_united

if(!require('igraph'))install.packages('igraph')

## Loading required package: igraph

## 
## Attaching package: 'igraph'

## The following object is masked from 'package:tidyr':
## 
##     crossing

## The following objects are masked from 'package:dplyr':
## 
##     as_data_frame, groups, union

## The following objects are masked from 'package:stats':
## 
##     decompose, spectrum

## The following object is masked from 'package:base':
## 
##     union

library(igraph)
bigram_counts <- bigrams_separated %>% 
  count(word1, word2, sort = TRUE) 
bigram_counts

bigram_graph <- bigram_counts %>%
  graph_from_data_frame()
bigram_graph

## IGRAPH b0f7236 DN-- 20 16 -- 
## + attr: name (v/c), n (e/n)
## + edges from b0f7236 (vertex names):
##  [1] an      ->obedient  fan     ->has       has     ->stopped  
##  [4] have    ->lost      hope    ->in        i       ->have     
##  [7] in      ->my        is      ->an        looks   ->beautiful
## [10] lost    ->hope      my      ->life      obedient->daughter 
## [13] rekha   ->is        riya    ->looks     stopped ->moving   
## [16] the     ->fan

if(!require('ggraph'))install.packages('ggraph')

## Loading required package: ggraph

library(ggraph)
set.seed(222)

ggraph(bigram_graph, layout = "kk") +
  geom_edge_link() +
  geom_node_point() +
  geom_node_text(aes(label = name), vjust = 1, hjust = 0.5)

set.seed(2020)

a <- grid::arrow(type = "closed", length = unit(.15, "inches"))


ggraph(bigram_graph, layout = "kk") +
  geom_edge_link(aes(edge_alpha = n), show.legend = FALSE,
                 arrow = a, end_cap = circle(.07, 'inches')) +
  geom_node_point(color = "lightblue", size = 5) +
  geom_node_text(aes(label = name), vjust = 1.5, hjust = 0.5) +
  theme_void()

Sentiment Analysis using Sentimentr Package/Sentiment Analysis - 1

Get sentiment of the sentence using function sentiment() to understand polarity of text(if it is a positive or negative)

Get aggregate of sentiment usign sentiment_by() function to understand the total polarity of text

Get emotion of the sentence using function emotion() to understand how the writer/author of sentence feels based on text

Detect profanity if text using function profanity()

Get Sentences from a given text as list using get_sentences() method

extract_sentiment_terms

Highlighted Sentences Based on Sentiment

Plot the sentiment analysis