Understanding Manipulative Actions And Political Language On Twitter: Exploring Trending Topics And The 2023 Turkish Presidential Election

Official URL: Includes bibliographical references (leaves. 52-55).

This thesis examines coordinated activities and reflections of public events via sentiment scores on Twitter, in the context of trending topics in Turkey and tweets sent related to Turkey’s presidential election in 2023. There are two objectives of this study. The first one is to understand how manipulative actions take place on Twitter, especially in trending topics, and to check whether similar actions occur in election-related discussions. The second one is to examine how public events and the political language used by presidential candidates are reflected in Twitter data, especially during the election period. Our analysis encompassed two datasets, consisting of tweets related to trending topics and tweets specifically mentioning the presidential candidates. To identify manipulative actions, we leveraged the suspension status of each tweet. Additionally, in the sentiment analysis part, we examined the sentiment scores of tweets based on various factors, including their types, suspension status, mentioned candidates, and the followers’ status of the respective candidates. Through our comprehensive analysis, we offer profound insights into the impact of political events and electoral outcomes on the aforementioned dimensions, thereby enhancing our understanding and providing valuable insights into the utilization of political language.