One of the major projects I have worked on till now outside of curriculum is a Twitter Government Sentiment Analysis. It is not just a regular sentiment analysis from a tweet input but has a lot more functionalities and complexity. To give a brief idea about what it does, the project searches a hashtag and displays real time tweets, the user who tweeted it, total retweet count of that tweet, all the hashtags used in each tweet, and most importantly the sentiment analysis of each tweet (whether it is a positive tweet or negative). The result shows the most recent 200 tweets from the day you want it to be searched from by taking a hashtag and date as input from the user. At the top of the result table, you get the total positive tweets percentage and negative tweets percentage of that hashtag. It is a full-fledged website with attractive frontend and smooth backend developed by me. I have developed the sentiment analysis model using logistic regression algorithm, and sqlite3 for database management. The major libraries I needed in the machine learning part are sklearn for logistic regression, nltk for preprocessing and tweepy for twitter authentication and tweets handling. I used matplotlib and seaborn libraries for result visualization to improve the accuracy of my project. The final accuracy I achieved is 98%. Coming to the website building, I have used Flask as my backend language and HTML, CSS, Javascript for frontend. Using Javascript, I was able to add beautiful scroll-animation effect to my project which gave it a more subtle and pleasing user experience. This project can be very useful for companies wanting to take a quick review on what's being said about their product on social media, especially from a specific period where they have made a significant change in their servicing or any other prospect of their product. They can understand the percentage of people who find their product/service positive or negative within seconds.