Problem statement: Build a model that classifies tweets whether they are positive, negative or neutral sentiments. Metric of sucess Build a model with above 80% accuracy that can classify tweets into negative, positive and neutral sentiments. Understanding the Context Moringa school wants to improve the interaction experience with their users on Twitter. Optimising on twitter interaction is important because users can be, consumers, potential consumers,business parters or potential business parters. By classifying the tweets we can help the relevant department addressing emerging and pressing issues affecting the users Experimental Design Scraping for tweets from twitter Basic clean( removing puctuations, username, hashtags, url links) labeling of tweets Undertaking Exploratory data analysis Sentimentetting up parameters for the neural network Build a baseline model. Improve the model, applying transfer learning.