Project on YouTube regarding predicting clickbait and metadata
Many social media platforms have an indicator if a post is potentially false information and/or misleading. YouTube is no exception as the largest social media platform. This project aims to determine if the prediction of a YouTube video's content is clickbait or not based mainly on the titles of the video as well as several metadata statistics. This uses a sentiment analysis and a Naive Bayes classification model to discover this. This projects also determines if the number of dislikes has grown or decreased since the removal of the dislike count in November 2021. This uses several regression techniques and statistical testing methods.
The report is 94 pages with many graphics and the code at the end. Follow along with the table of contents for precise navigation of specific sections.