Tweets about the products we are tracking are aggregated and de-duplicated from the twitter streaming api. The tweets collected from this api are then bundled and sent into a Kafka server. The Kafka server holds onto the messages until the sentiment analysis pipeline is ready to analyze. The sentiment analysis is preformed on a Hadoop cluster using deep models and a pre-computed sentiment tree-bank. The sentiment scores for tracked products are then updated in a mongodb and displayed with nodejs.