Analyzed over 8,000 Twitter Messages to calculate scores based on how positive/negative the words are; Extracted features from Twitter Messages, removed irrelevant stopwords/hyperlinks/illegal characters using Python script to give polarity scores; Integrated with Python open source APIs like Tweepy, TextBlob, NLTK to achieve more precise analysis; Testing the accuracy of results using Weka classification which based on SMO, J48, and Naive Bay. The best accuracy was 83% Read academic research papers to get a better understanding of sentiment analysis, and different algorithms to calculate polarity scores;
