This repo is about the tweet analysis tutorial we learnt on seminar. Here is the original tutorial: https://dev.to/rodolfoferro/sentiment-analysis-on-trumpss-tweets-using-python- Many thanks to the author!
Other than replicating the tutorial above, I've tried to test the accuracy of sentiment analysis. First I choose a tweet from Federer labeled as "negative":"#Fbf to Match for Africa 4 Seattle last month! What an unbelievable event- thanks to everyone who made it possible!" - to me this is not a negative tweet.
I removed the word "unbelievable" and then test it again, the sentiment analysis has changed from "negative" to "positive".