The insurgency in Southern Thailand has been a protracted and complex conflict, drawing international attention due to its impact on regional stability and human rights. Media coverage plays a crucial role in shaping public perception and understanding of such conflicts. This project aims to analyze the historical coverage of Southern Thailand's insurgency by The Guardian using web scraping and text analysis techniques. The research question guiding this analysis is: "What are the dominant sentiments and topics associated with the Southern Thailand insurgency portrayed in The Guardian over the past two decades?" By extracting and processing articles spanning nearly two decades, the project provides insights into the prominent topics (word frequency data), the tone of articles (sentiment scores), and the thematic content and structure (topic modeling data) of The Guardian's coverage. The findings from this analysis will contribute to a deeper understanding of how international media portrays long-standing regional conflicts.
This project utilizes news articles from The Guardian to analyze the coverage of the Southern Thailand insurgency. The dataset consists of 25 news articles selected through web scraping, focusing on the period from 2004 to 2021. The choice of this timeframe allows for a comprehensive examination of how media narratives and reporting on the conflict have evolved over nearly two decades.
| Data Source | The Guardian | | --- | --- | | Quantity | 25 news articles | | Timeframe | 2004 to 2021 | | Search Terms | "southern Thailand conflict," "southern Thailand insurgency," "southern Thailand peace," "Pattani conflict," "Yala conflict," "Narathiwat conflict," "Songkhla conflict" |
The articles were chosen based on their relevance to the Southern Thailand insurgency, ensuring a diverse representation of topics and sentiments. The dataset encompasses a variety of articles, including news reports, feature stories, and opinion pieces, which collectively provide a broad perspective on the media portrayal of the conflict.
The selected articles were scraped and processed to extract relevant text data for analysis. This dataset will be used to explore the dominant themes, sentiments, and the narrative surrounding the Southern Thailand insurgency as reported by The Guardian.
### Step 1: Load Required Libraries ###
# Install the packages
install.packages(c("rvest", "dplyr", "stringr", "tm", "tidytext", "ggplot2", "tidyr", "topicmodels
", "SnowballC"))
# Load the libraries
library(rvest)
library(dplyr)
library(stringr)
library(tm)
library(tidytext)
library(ggplot2)
library(tidyr)
library(topicmodels)
library(SnowballC)
### Step 2: Scrape the Web Pages ###
# Function to scrape an article
scrape_article <- function(url) {
webpage <- read_html(url)
# Extract title
title <- webpage %>%
html_node("title") %>%
html_text(trim = TRUE)
# Extract date
date <- webpage %>%
html_node("meta[property='article:published_time']") %>%
html_attr("content") %>%
str_extract("\\\\d{4}-\\\\d{2}-\\\\d{2}") # Extract date format
# Extract article content
content <- webpage %>%
html_nodes("p") %>%
html_text(trim = TRUE) %>%
paste(collapse = " ")
# Return as a data frame
return(data.frame(title = title, date = date, content = content, url = url, stringsAsFactors = FALSE))
}
# Apply the function to all URLs
urls <- c(
"<https://www.theguardian.com/world/2004/jul/25/jasonburke.theobserver>",
"<https://www.theguardian.com/world/2004/oct/27/thailand>",
"<https://www.theguardian.com/world/2004/oct/26/thailand?CMP=gu_com>",
"<https://www.theguardian.com/world/2004/may/06/worlddispatch.johnaglionby>",
"<https://www.theguardian.com/theguardian/2005/feb/28/theeditorpressreview.thailand>",
"<https://www.theguardian.com/world/2005/mar/28/thailand>",
"<https://www.theguardian.com/world/2006/sep/04/thailand>",
"<https://www.theguardian.com/world/2006/jan/28/comment.simontisdall>",
"<https://www.theguardian.com/world/2006/sep/21/thailand2>",
"<https://www.theguardian.com/world/2008/jan/14/thailand.ianmackinnon>",
"<https://www.theguardian.com/commentisfree/2009/dec/01/thailand-malay-muslims>",
"<https://www.theguardian.com/world/2013/feb/28/thailand-muslim-separatists-peace-talks>",
"<https://www.theguardian.com/world/2015/may/15/string-of-bomb-attacks-in-thailands-restive-south-injure-18-people>",
"<https://www.theguardian.com/world/2015/jul/11/seven-dead-and-dozens-injured-in-attacks-in-thailands-south>",
"<https://www.theguardian.com/world/2015/apr/11/koh-samui-car-bomb-seven-people-hurt-on-thai-tourist-island>",
"<https://www.theguardian.com/world/2015/nov/13/four-die-in-southern-thailand-bombing>",
"<https://www.theguardian.com/world/2015/aug/18/bangkok-bomb-not-in-keeping-with-insurgent-attacks-says-army-chief>",
"<https://www.theguardian.com/global-development/2016/aug/24/bridging-language-divide-thailand-muslim-patani-malay-schools-ethnic-tension>",
"<https://www.theguardian.com/world/2016/aug/24/two-bombs-explode-in-thai-beach-resort-of-pattani>",
"<https://www.theguardian.com/world/2016/aug/19/thai-police-hunt-one-man-confusing-inquiry-into-bombings>",
"<https://www.theguardian.com/world/2016/aug/11/one-dead-and-20-injured-in-two-bomb-blasts-in-hua-hin-thailand>",
"<https://www.theguardian.com/world/live/2016/aug/12/thailand-explosions-resort-towns-phuket-hua-hin?filterKeyEvents=false&page=with:block-57ad4e6fe4b024cdec3d6c41>",
"<https://www.theguardian.com/world/2016/aug/12/thailand-attacks-who-are-likely-perpetrators>",
"<https://www.theguardian.com/world/2019/nov/06/dead-killed-in-biggest-attack-in-thailands-restive-south-in-years>",
"<https://www.theguardian.com/global-development/2021/aug/17/the-woman-on-a-mission-to-expose-torture-in-thailands-troubled-south>"
)
# Scrape all articles
articles <- do.call(rbind, lapply(urls, scrape_article))
# View the scraped data
View(articles)
### Step 3: Data Preprocessing ###
# Function to clean and preprocess text
clean_text <- function(text) {
text <- tolower(text) # Convert to lowercase
text <- removePunctuation(text) # Remove punctuation
text <- removeNumbers(text) # Remove numbers
text <- removeWords(text, stopwords("en")) # Remove stopwords
text <- wordStem(text, language = "en") # Perform stemming
text <- stripWhitespace(text) # Remove extra whitespace
return(text)
}
# Apply the cleaning function to the article content
articles$content_clean <- sapply(articles$content, clean_text)
# View the cleaned content
head(articles$content_clean)
### Step 4: Word Frequency ###
# Tokenize the cleaned content
articles_tidy <- articles %>%
unnest_tokens(word, content_clean)
# Count word frequencies
word_freq <- articles_tidy %>%
count(word, sort = TRUE)
# View the most frequent words
head(word_freq, 22)
# Visualize the most common words
word_freq %>%
filter(n > 30) %>%
ggplot(aes(x = reorder(word, n), y = n)) +
geom_col() +
coord_flip() +
labs(title = "Most Common Words in Articles", x = "Word", y = "Frequency")
### Step 5: Sentiment Analysis ###
# Load sentiment lexicon
sentiment_lexicon <- get_sentiments("bing")
# Perform sentiment analysis
article_sentiments <- articles_tidy %>%
inner_join(sentiment_lexicon, by = "word") %>%
count(url, sentiment) %>%
pivot_wider(names_from = sentiment, values_from = n, values_fill = list(n = 0)) %>%
mutate(sentiment_score = positive - negative)
# View sentiment scores by article
article_sentiments
# Visualize sentiment scores
article_sentiments %>%
ggplot(aes(x = url, y = sentiment_score, fill = factor(sentiment_score > 0))) +
geom_col() +
coord_flip() +
labs(title = "Sentiment Scores by Article", x = "Article", y = "Sentiment Score")
### Step 6: Topic Modeling ###
# Create a document-term matrix
dtm <- articles_tidy %>%
count(url, word) %>%
cast_dtm(url, word, n)
# Fit the LDA model with a chosen number of topics (k = 3)
lda_model <- LDA(dtm, k = 3, control = list(seed = 1234))
# Extract topics
topics <- tidy(lda_model, matrix = "beta")
# View top terms in each topic
top_terms <- topics %>%
group_by(topic) %>%
top_n(10, beta) %>%
ungroup() %>%
arrange(topic, -beta)
View(top_terms)
# Visualize top terms in each topic
top_terms %>%
ggplot(aes(x = reorder_within(term, beta, topic), y = beta, fill = factor(topic))) +
geom_col(show.legend = FALSE) +
facet_wrap(~ topic, scales = "free") +
coord_flip() +
scale_x_reordered() +
labs(title = "Top Terms in Each Topic", x = "Term", y = "Beta")
The word frequency analysis of the collected news articles provides insights into the most prominent topics and recurring themes related to the Southern Thailand insurgency. The analysis focuses on the most frequent words, which help highlight key subjects discussed in the articles.