Inspiration
Youtube algorithm. The same way Youtube is able to generate a short feed a user should be able to see how their previous history is impacting future recommendation
What it does
Labels shorts at run time.
How we built it
https://github.com/jonss0777/HackHavard
Challenges we ran into
Finding a way to retrieve Youtube short’s description information. Options include using the YouTube Data API for video details or employing web scraping if the API doesn’t suffice. Understanding YouTube's HTML structure is crucial, along with adhering to any usage restrictions.
Model user behavior: analyzes how users interact with YouTube Shorts, focusing on metrics like viewing duration, engagement (likes, comments, shares), and demographics. By applying data analytics and machine learning, trends can be identified to inform content recommendations and enhance user experience.
Training labeling algorithm: involves developing a machine learning model to categorize YouTube Shorts by content type. It begins with compiling a labeled dataset for supervised learning, allowing the algorithm to improve its classification accuracy. The goal is to automate labeling, making content easier to organize and discover.
Accomplishments that we're proud of
Facing the different challenges involved in creating a chrome extension.
What we learned
How chrome extensions work. How to manipulate large html documents
What's next for mindshort
Continue to build on our ai model. Include more parameters and create a website where the user can visualize the create content data.
Log in or sign up for Devpost to join the conversation.