ANLY540 - Analysis of Human Language - Executive Session 3

Objective

The research study intends to analyze the textual differences among 30 chapters from the book, ‘Thirty Strange Stories’ written by H.G Wells. Each of the 30 chapters from this book represents an individual story. Analysis will include assessing the text for verbiage and colexemes, sentiment analysis and topic modeling/clustering. Given that the 30 different stories are written by the same author, study intends to understand the relationship between word usage and storyline.

Introduction

The author of Thirty Strange Stories is H.G. Wells, who is well known for writing books such as the War of the worlds, The Invisible Man and The Time Machine. The text for the book has been sourced from the Gutenberg corpus.

Below is a list of the 30 stories covered in this book:

• The Strange Orchid
• Æpyornis Island
• The Plattner Story
• The Argonauts Of The Air
• The Story Of The Late Mr. Elvesham
• The Stolen Bacillus
• The Red Room
• A Moth (Genus Unknown)
• In The Abyss
• Under The Knife
• The Reconciliation
• A Slip Under The Microscope
• In The Avu Observatory
• The Triumphs Of A Taxidermist
• A Deal In Ostriches
• The Rajah’s Treasure
• The Story Of Davidson’s Eyes
• The Cone
• The Purple Pileus
• A Catastrophe
• Le Mari Terrible
• The Apple
• The Sad Story Of A Dramatic Critic
• The Jilting Of Jane
• The Lost Inheritance
• Pollock And The Porroh Man
• The Sea Raiders
• In The Modern Vein
• The Lord Of The Dynamos
• The Treasure In The Forest

Relevant literture pertaining the study of natural language processing through textual analysis of various authors’ works will be reviewed.

Hypothesis / Problem Statement

Statistical Analysis Plan

The research study intends to touch upon the following analysis

• Sentiment Analysis
• Basic exploratory data analysis: Dispersion and Location
• Topics Models
• Word Frequency
• Co Lexemes analysis
• Attraction and Reliance
• PMI and Odds Ratio

Method

library(gutenbergr)
library(stringr)
library(dplyr)
library(tidyr)
library(tm)
library(topicmodels)
library(tidyverse)
library(tidytext)
library(slam)
library(ggplot2)
library(wordcloud)

# download the entire book
data <- gutenberg_download(59774)

# Remove the first and last unwanted rows
data <- data[96:12231,]

# check for UPPER case 
data$check <- data$text == toupper(data$text)

# filter out empty rows
data <- data %>% filter(text != "" )

# create row number
data <- data %>% mutate(row_num = row_number())

# remove incorrectly detected chapter headings
data <- data[-c(205,2823,3291,4194,4630,4631,5833,5923,5975,6064,6109,8864,8989,9137,9205),]

# Create Chapter
data$chapter <- cumsum(data$check)

# Create a separate datset for chapter 
chapters_headings <- filter(data, check == TRUE) %>% rename(chapter_name = text) %>% 
                      select(chapter, chapter_name) 


# Clean up and join with chapter headings 
data <- data %>% mutate(title = "Thirty Strange Stories") %>% mutate(row_num = row_number()) %>% 
        select(title, text, row_num, chapter) %>% left_join(chapters_headings)

# Remove leading and trailing white spaces
data <- data.frame(lapply(data, trimws), stringsAsFactors = FALSE)
data$row_num <- as.integer(data$row_num)
data$chapter <- as.integer(data$chapter)

# Words for initial analysis
words <- data %>% unnest_tokens(word, text)

# Remove leading and trailing white spaces from chapters_heading for later use
chapters_headings <- data.frame(lapply(chapters_headings, trimws), stringsAsFactors = FALSE)

words %>% filter(chapter == 1) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 2) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 3) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 4) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 5) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 6) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 7) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 8) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 9) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 10) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 11) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 12) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 13) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 14) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 15) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 16) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 17) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 18) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 19) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 20) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 21) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 22) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 23) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 24) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 25) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 26) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 27) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 28) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 29) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

words %>% filter(chapter == 30) %>%
  anti_join(stop_words) %>% 
  count(word) %>%
  with(wordcloud(word, n, max.words = 50,))

sentiment_books <- words %>%
  inner_join(get_sentiments("bing")) %>%
  count(chapter_name, index = row_num %/% 20, sentiment) %>%
  spread(sentiment, n, fill = 0) %>%
  mutate(sentiment = positive - negative)

sentiment_books_1 <- sentiment_books %>% inner_join(chapters_headings[1:6,])
sentiment_books_2 <- sentiment_books %>% inner_join(chapters_headings[7:12,])
sentiment_books_3 <- sentiment_books %>% inner_join(chapters_headings[13:18,])
sentiment_books_4 <- sentiment_books %>% inner_join(chapters_headings[19:24,])
sentiment_books_5 <- sentiment_books %>% inner_join(chapters_headings[25:30,])

ggplot(sentiment_books_1, aes(index, sentiment, fill = chapter_name)) +
  geom_col(show.legend = FALSE) +
  facet_wrap(~chapter_name, ncol = 2, scales = "free_x") +
  theme_bw()

ggplot(sentiment_books_2, aes(index, sentiment, fill = chapter_name)) +
  geom_col(show.legend = FALSE) +
  facet_wrap(~chapter_name, ncol = 2, scales = "free_x") +
  theme_bw()

ggplot(sentiment_books_3, aes(index, sentiment, fill = chapter_name)) +
  geom_col(show.legend = FALSE) +
  facet_wrap(~chapter_name, ncol = 2, scales = "free_x") +
  theme_bw()

ggplot(sentiment_books_4, aes(index, sentiment, fill = chapter_name)) +
  geom_col(show.legend = FALSE) +
  facet_wrap(~chapter_name, ncol = 2, scales = "free_x") +
  theme_bw()

ggplot(sentiment_books_5, aes(index, sentiment, fill = chapter_name)) +
  geom_col(show.legend = FALSE) +
  facet_wrap(~chapter_name, ncol = 2, scales = "free_x") +
  theme_bw()