Improving Youtube's Ad-Revenue Algorithm for Sex Educators
For this project, I tried to improve Youtube's ad-revenue algorithm so that it did not disproportionately affect sex education content creators on the platform. This was accomplished by utilizing an audio and text based approach to build a classification algorithm that could distinguish between educational and explicit content. The model itself was then deployed onto a Flask app that had an interface through a local Google Chrome extension.
Repository Structure
- Flask: Flask app that I built utilizing my model
- Notebooks: Data Scraping, EDA, and Modeling done on respective Jupyter notebooks
- chrome: Google Chrome extension that was built to communicate with the Flask app
- kojak_MVP.md: Document describing my minimum viable product for this project.