Performance Prediction of Product/Person Using Real Time Twitter Tweets

Over a previous decade people have experienced an exponential boom in the usage of online resources in specific social media and microblogging internet site such as Twitter, Facebook, Instagram and YouTube. Many businesses and agencies has identified these sources as a wealthy mine of marketing information. On such platforms, massive quantities of records are produced (e.g.: 5000 tweets per 2d on twitter), this representing an chance for companies to check their social impact and people opinions towards their products, and even frequent people can additionally discover out what is a performance of a certain product or the overall performance of a particular political personality. In this project, we fetch the given number of tweets from users and classify it as Positive, Negative and Neutral by the usage of supervised machine learning approach. In this method we’re analyzing the Polarity and Subjectivity of the tweets and then later we’re using NLP to classify the raw records into records body which gets rid of the undesirable words from each of the tweets. Neutral words like ‘as, the, of’ are eliminated from the tweets. Using NLP, we get better results of the tweets, later we classify the tweets using classifying algorithms like Random Forest Classifier, Decision Tree Classifier, Logistic Regression, and Support Vector Classifier. Later it compares the result of tweets which had been analyzed before processing into NLP. We are also using Data Visualization for phrase frequencies, and for displaying a pie or bar chart of a variety of positive, negative and impartial tweets.