With the advent of modern technology and innovations, the entire world is undergoing a radical transformation. The Internet has become essential for everyone, and the world wide web is used in almost every field. Affordable access to web content and mobile and other devices allows most people to participate actively in and review various activities on the internet. Human life is full of emotions and opinions. People communicate with their fellow beings using opinions and emotions over various issues. Sentiment analysis or opinion mining identifies the emotions expressed in texts on a varied range of subjects. The field of sentiment analysis has paved the way for easy preprocessing of reviews for analysis and using the proper classification algorithms to validate the usage of the classifier and predict the overall nature of sentiment expressed on topics like politics, products, movies, and other daily social problems. Twitter has grown in popularity as a popular micro-blog application where customers can express themselves. Twitter data investigation and analysis is a field in which the research community and industry have given more attention during the last decade. Tweets are analyzed and the polarity of the content of the expression is studied to know the intricacies of the emotions of people over various subjects. As a result, this research paper investigates sentiment classification algorithms: The decision tree and Naïve Bayesian classifiers on Twitter data and their outcomes. We have used a dynamic data set of tweets using Twitter API calls with 200 tweets per call. In this Textblob is used for calculating the polarity. The experiments are carried out using different classifiers and achieved good performance. We have been fortunate to achieve 90.79% accuracy and the F1-Score was 0.88 using Naïve Bayes classifier on this Twitter data stream.

API, Twitter analysis, Sentiment analysis, Descriptive Statistical Analysis