Analyzing Public Sentiment: A Deep Dive into Twitter Discourse on the 2022 No Confidence Motion in Pakistan

Twitter, among other widely used social media platforms, is a prominent space for the global expression of feelings and opinions. In Pakistan, political parties actively utilize Twitter to connect with the public, fostering a continuous exchange of viewpoints. This results in a substantial daily volume of millions of tweets covering diverse political topics. This study conducts sentiment analysis on tweets gathered from the X platform (formerly Twitter) concerning a significant political event in Pakistan during the year 2022. The event was related to the No Confidence Motion which resulted in the dismissal of sitting Prime Minister Imran Khan. This work prepares dataset which comprises of 10000 tweets and is publicly available on GitHub. Tweets are classified as positive, negative, or neutral by using machine learning, deep learning, and transformers. Then, the comparison is made between these approaches using benchmark evaluation measures such as recall, precision, and F1-score. Multilingual BERT, a deep learning approach, turns out to be the best classifier for the collected tweets with the highest F1-score of 77.27%.