SENSITIVE ANALYSIS OF POLITICAL TWITTER USAGE IN VOTING FOR THE PUBLIC

Abstract

Classical political analysis tasks include surveying the public to gauge their opinions in order to make predictions about future elections (both national and local). But surveys may be costly to do, and if the sample isn't representative of the community, the findings might be skewed. We provide two novel approaches to using the findings of tweet sentiment analysis in survey research. Our first approach only relies on the sentiment analysis findings from tweets, which surpasses other widely-recognized methodologies. We also present a new hybrid method that estimates election outcomes using public opinion surveys and Twitter. The outcome of the 2023 Greek national election may be predicted using this approach, which allows for more precise, frequent, and cost-effective public opinion prediction.
New opportunities for public opinion estimate using social media platforms were opened up by our technique, which showed less variance from the actual election outcomes than traditional public opinion surveys.
Terms related to this index include: political tweet analysis, popularity score, sentiment analysis, and Twitter data.