Bigram Analysis

library(gutenbergr)
library(tidyr)
library(tidytext)
library(dplyr)
library(ggplot2)
library(knitr)

gutenberg_works(title == "The Time Machine") # extract The Time Machine

## # A tibble: 1 x 8
##   gutenberg_id title  author   gutenberg_author… language gutenberg_books…
##          <int> <chr>  <chr>                <int> <chr>    <chr>           
## 1           35 The T… Wells, …                30 en       Science Fiction…
## # ... with 2 more variables: rights <chr>, has_text <lgl>

time_machine <- gutenberg_download(c(35), meta_fields = "title") # pass gutenberg_id

machine <- time_machine$text[grep("Morlocks", time_machine$text)]
length(machine)

## [1] 46

head(machine, 8)

## [1] "I had seen of the Morlocks--that, by the by, was the name by which"  
## [2] "'Then came troublesome doubts. Why had the Morlocks taken my Time"   
## [3] "abruptly to trouble about the Morlocks, and was only concerned in"   
## [4] "influence of the Eloi, whose disgust of the Morlocks I now began"    
## [5] "spectral Morlocks sheltered from the glare. The place, by the by,"   
## [6] "Morlocks at any rate were carnivorous! Even at the time, I remember"
## [7] "blackness I could hear the Morlocks rustling like wind among leaves,"
## [8] "Morlocks and was speedily clambering up the shaft, while they stayed"

bigrams <- time_machine %>%
  unnest_tokens(bigram, text, token = "ngrams", n = 2)

bigrams_separate <- bigrams %>%
  separate(bigram, c("word_1", "word_2"), sep = " ")

head(bigrams_separate)

## # A tibble: 6 x 4
##   gutenberg_id title            word_1  word_2
##          <int> <chr>            <chr>   <chr>  
## 1           35 The Time Machine the     time   
## 2           35 The Time Machine time    machine
## 3           35 The Time Machine machine by     
## 4           35 The Time Machine by      h      
## 5           35 The Time Machine h       g      
## 6           35 The Time Machine g       wells

bigrams_filter <- bigrams_separate %>%
  filter(!word_1 %in% stop_words$word) %>%
  filter(!word_2 %in% stop_words$word) %>%
  unite(bigram, word_1, word_2, sep = " ")

head(bigrams_filter)

## # A tibble: 6 x 3
##   gutenberg_id title            bigram          
##          <int> <chr>            <chr>           
## 1           35 The Time Machine time machine    
## 2           35 The Time Machine time traveller  
## 3           35 The Time Machine recondite matter
## 4           35 The Time Machine grey eyes       
## 5           35 The Time Machine eyes shone      
## 6           35 The Time Machine fire burned

bigrams_frequent <- bigrams_filter %>% dplyr::count(bigram) %>% dplyr::arrange(desc(n))
bigrams_frequent2 <- bigrams_frequent %>% inner_join(bigrams_filter)

## Joining, by = "bigram"

bigrams_frequent2 <- as.data.frame(unique(bigrams_frequent2))
head(bigrams_frequent2, 10)

##              bigram  n gutenberg_id            title
## 1    time traveller 54           35 The Time Machine
## 2      time machine 39           35 The Time Machine
## 3      white sphinx 10           35 The Time Machine
## 4   green porcelain  8           35 The Time Machine
## 5      looked round  7           35 The Time Machine
## 6   time travelling  7           35 The Time Machine
## 7  time traveller's  6           35 The Time Machine
## 8       upper world  6           35 The Time Machine
## 9      world people  6           35 The Time Machine
## 10     bronze doors  4           35 The Time Machine

nature_bigrams <- bigrams_frequent2[grep("time", bigrams_frequent2$bigram), ]
nature_bigrams2 <- nature_bigrams %>% arrange(desc(n)) %>% head(15)
kable(nature_bigrams2)