Twitter Data Mining and Sentiment Analysis

Text Mining in Twitter Twitter API Tweepy NLTK Sentiment Analysis with VADER

Twitter API REST API Streaming API Persistent, realtime connection Tweets pushed as they occur Filters on hashtags or keywords Data mining or individual topics

Steps to Capture Tweets Tweepy Oauth Authentication Listener on_data on_error Save to database

Sentiment Analysis Aims to measure the emotional response of a writer, speaker or (in our case) tweeter Positive/Negative

VADER Assigns values for positive, negative, neutral and composite Values between -1 and 1 Trained on social media

How good is this tweet? Tweet Text Compound Positive Neutral Negative This is a bad tweet -0.5423 0.462 0.538 This is a good tweet 0.4404 0.516 0.484 This is a GOOD tweet! 0.6027 0.567 0.433 BEST TWEET EVVARRR!!!!! 0.7482 0.729 0.271