Exploring the Book of Esther in the Tidy Verse

November 30, 2017

library(httr)
library(tidytext)
library(jsonlite)
library(tidyverse)
library(wordcloud)
library(XML)
source("~/keys.R")
key <- biblia

There is much to be learned in the form of writing and the relationship it has within and outside the text. The question we might want to ask is there a way to identify the story arc of the situation. Using spacy to tag all of the words will increase the likelyhood to better understand the

style <- "oneVersePerLineFullReference" #oneVersePerLine, bibleTextOnly, oneVersePerLineFullReference
output <- "html"
passage <-"Esther1-10"
resp <- GET(paste0("https://api.biblia.com/v1/bible/content/LEB.",output,"?passage=",passage,"&style=",style,"&key=",key))

# parse html
html <- content(resp,"text")
doc = htmlParse(html, asText=TRUE)
plain.text <- xpathSApply(doc, "//p", xmlValue)
plain.text <-  gsub('[\r\n\t]', '', plain.text)
plain.text <- tibble(Text = plain.text)

Create the Data Frame

TODO: Need to make sure to figure out the book regex to include books such as 1 Timothy where there are parts of the book

passage <- plain.text %>% 
  mutate(
        Book = str_extract(Text, "^([*?\\w]+)"), 
        Chapter = as.numeric(str_extract(Text,"\\d+(?=:\\n?)")), 
        Verse = as.numeric(str_extract(Text,"(?<=\\:)\\d+")),  
        Line = str_replace(Text, "^([*?\\w]+\\s\\w+\\:\\w+)","")) 

tibble(Text = passage$Text)

data("stop_words") # Remove stop words 
passage_frame <- passage %>% select(Book, Chapter, Verse, Line) %>% unnest_tokens(word, Line) %>%
    anti_join(stop_words)
head(passage_frame)

Here to simply look at the entire passage as a whole, this is what we are able to retrieve in counts after having removed the words.

passage_frame_cnt <- passage_frame %>% count(word, sort = TRUE)
head(passage_frame_cnt)

passage_framea_cnt <- passage_frame %>% group_by(Chapter) %>% count(word, sort = TRUE)
head(passage_framea_cnt)

Word frequencies by chapter

library(ggplot2)

g <- passage_framea_cnt %>%
  top_n(5) %>%
  filter(n > 1) %>%
  ggplot(aes(reorder(word, n), n, fill = as.factor(Chapter))) +
  geom_col() +
  xlab(NULL) +
  facet_grid(.~Chapter) + 
  coord_flip()

g

gb <- passage_framea_cnt %>% 
  filter(word %in% c("king","jews","mordecai","haman","esther", "women")) %>%  
  ggplot(aes(Chapter, n, fill = word )) + 
  geom_col() + 
  labs(x = "Chapter", y = "Occurance") +
  facet_grid(. ~ word)
gb
  
ga <- passage_framea_cnt %>% filter(word %in% c("king","jews")) %>%  
  ggplot(aes(Chapter, n )) + 
  geom_col() + 
  labs(x = "Chapter", y = "Occurance") +
  scale_fill_brewer() + 
  facet_grid(. ~ word) 
  

  ga

Sentiment Analysis of the story and its progression

Generating the Wordcloud for each chapter

  passage_frame_cnt %>%
  with(wordcloud(word, n, max.words = 100, random.color = TRUE))

###References

Timeseries Analysis Using R and Advantager Words With Friends Musings