A Social Analysis of Thailand's 2023 Election Through Twitter Feeds

Nowadays, it cannot be denied that social media has been in many parts of people's lives, especially in entertainment, business and education. Public figures including politicians frequently use social platforms to connect with their audience, for example Facebook, Instagram, Tiktok and etc. One of the most commonly used is a platform called Twitter or also known as X. It is a platform that people can share their opinions in short messages. Since there is an alphabet limitation, people can deliver and receive message rapidly. It is a perfect platform that we could use to update the real-time situations for example election, emergency alert, natural disasters and etc. According to its characteristic, it can be used as a data source for social media analytics, which involves collecting and analyzing data to perform business decisions. The social media analytics has become commonplace because the number of people using social media tends to increase every second. Specifically, Twitter sentiment analysis offers a quick and inexpensive tool for monitoring and predictions. In this work, Twitter is used as a data source to monitor and predict Thailand Election in 2023 because this election highly attracts attention from Thais and international. Moreover, Twitter is a significant platform that is used to debate about political events in Thailand. The result of this election potentially indicates Thailand's future, including economics, public health, environment, international relations and etc. Our work performs sentiment analysis to categorize the data and mainly focuses on extracting sentiment from text without considering semantic word relations and co-occurrences because it is appropriate and functional for Thai sentence analysis. Based on the result, it can be seen that Twitter plays a significant role in politics, with politicians wielding considerable influence over public perception. Comparing of the actual election result and the predicted result, this analysis demonstrates the potential for prediction precisely. The candidates with high positive sentiments have higher possibility to win the election while those with low sentiments has lower possibility.