US presidential election 2020 prediction based on Twitter data using lexicon-based sentiment analysis

Sentiment analysis is a technique of analyzing public opinion on a problem. The presidential election in the United States is a hot issue that will affect various aspects of the world. The goal of this analysis is to forecast the outcome of the US presidential election and to compare these results with the actual results of the polls. The sentiment analysis used in this study is the lexicon-based sentiment analysis. The method used in this research is data collection, data preprocessing, data mapping and sentiment analysis. The data in this study were obtained from Twitter taken one week before the United States presidential election was held. The model used in this research is V A D E R sentiment analysis. The data cleaning mechanism in this study uses a method in text mining, where the data is first cleaned of various things that are not considered important in the analysis. Furthermore, the data that can be used as material for analysis is saved again to make it easier to read the data. In the analysis, tweets from users are mapped and counted by the state of the United States of America. The result of this research is a prediction for the Democratic Party to win 22 votes over the Republican Party which received 19 votes. The results from the B B C show that the Democratic Party won with 24 votes, and the Republican Party only got 20 votes. With these results, the V A D E R sentiment analysis model can produce predictions following the actual results of the US presidential election. © 2021 IEEE.