Sentilyzer : Lexicon based Sentiment Analysis of Tweets

Abstract

Sentiment analysis is a process of deriving emotions from texts. Twitter Sentiment Analysis is used to find market insights and trends or changes in people's perceptions contained in tweets mentioning particular product or service on Twitter which is a popular social media platform. As the world is slowly shifting towards digital currency, we have seen that social media is critical in establishing public perception about cryptocurrency and crypto market trends. The purpose of this study is to determine if there is any correlation between positive or negative perceptions of cryptocurrencies and their value, and whether its popularity has changed over time. We have used lexicon-based approach for analyzing the sentiments of tweets referring to either of the two most popular cryptocurrency namely Bitcoin and Dogecoin. We extracted tweets from Twitter from different geographical locations and different time periods in order to make our sample size as diverse as possible so that we could get precise results. For visualization of these results, we have created dashboard in tableau. These visualisations show popularity of BTC and Doge through total count of tweets, their polarity i.e. negative, positive or neutral, sentiment scores, etc. Our study has shown that though there are times when Dogecoin’s value has significantly shot up due to positive tweets especially by most influential figures like Elon Musk but the effect isn’t permanent and popularity of Bitcoin and Dogecoin remains volatile. Keywords— lexicon, tweets, opinion, analysis, map, time, location, emotion, sentiment, score, polarity, visualisation, tableau, etc.