YouTube Comment Analyser

The ultimate tool for content creators, businesses, and viewers to easily analyse sentiments, emotions, and sarcasm in video comments. Take away the drudgery of crawling through endless comments!

Comment

Abstract

Are you a content creator? Are you a business looking to expand your brand online? Are you just an avid YouTube viewer who wants to save time? Our innovative tool, YouTube Comment Analyser will help all of you! Due to the exponential growth of online videos, understanding viewer sentiment, detecting sarcasm, and gauging emotional reactions in comments has become a daunting and a severely time-consuming task. YouTube Comment Analyser utilises cutting-edge Natural Language Processing techniques to automate this process effortlessly and efficiently.

Features

  • 3 Class Sentiment Analysis: YouTube Comment Analyser uses advanced natural language processing to evaluate the average sentiments in the comments section of a YouTube video. It provides a detailed breakdown of how positive, negative, or neutral comments are.
  • Emotion Analysis: YouTube Comment Analyser goes beyond simple sentiment analysis by categorising comments into emotions such as joy, anger, sadness, and fear. This provides a comprehensive emotional overview in lightning speed, crucial for tailoring content and engagement strategies.
  • Sarcasm Analysis: We all know audiences in YouTube videos are not the most serious commenters. Identifying this sarcasm can be extremely challenging, but YouTube Comment Analyser rises to the occasion with an advanced sarcasm detector, allowing creators and brands to understand their audience more effectively.
  • User-Friendly Interface: YouTube Comment Analyser is designed to be accessible to every kind of user. It can be interfaced through a user-friendly website, a popup extension, or an in-YouTube extension.

Challenges we ran into

  • NLP models were too slow (upwards of 2 minutes processing). We learnt how to optimise and quantise them, resulting in a <4s pipeline.
  • Handling and injecting into YouTube's interface proved difficult.
  • Learning how to manage safe and multiple web requests.

Built With

Share this project:

Updates