Motivation:
With the recent takeover of Twitter, there have been large amounts of discussion on the moderation of the platform and what messaging should, or should not be, allowed to be seen. Our project looks to automate and classify tweets as either positive, negative or neutral to aid in this discussion, and potentially provide a solution to the moderation of a large-scale social media platform like Twitter (X).
Problem Description:
Twitter, a significant social media platform, has a large quantity of active users generating large amounts of data daily. Out of these, a substantial portion includes tweets that carry a negative sentiment. These might be anything from negative statements to potentially harmful content like hate speech. The spread of such negative sentiment can have serious implications. It affects mental health, influences public opinion, and has the potential to spark action in the actual world. Understanding and identifying negative sentiment on Twitter is essential for moderation. It is not feasible to manually monitor and analyse tweets for negative sentiment because of the large volume of data. Additionally, human analysis is prone to bias, prejudice and inconsistent results. The main objective is to create a system that automatically identifies and classifies tweets according to their sentiment. The proposed solution involves the use of natural language processing for sentiment analysis to recognise and classify opinions expressed in text. NLP is well-suited for Twitter data analysis since it enables computers to interpret and categorise human language. This will be used to determine if the writer's sentiment is positive, negative, or neutral.