The idea is to create an index of youtube video metadata that can be sorted, searched and explored in any way you want. You start by scraping the page whether it be from the homepage, search results etc for all the video ids, and add them to a crawler that loads the pages and scrapes the metadata and more video ids and so on. And then just store it all to a database. Then make it into a website and a browser extension. The main motivation for doing this is because the algorithm is awful and this would offer a way for people to stop being controlled and actually see what they want. This could of course be done for any site that puts user content behind an algorithm but youtube is a fun example. Feel free to steal this idea.