YouTube Clickbait Detection 🎥🧐

Project Overview 📜

This project focuses on detecting clickbait content in YouTube videos by analyzing metadata, sentiment, and thumbnails. By utilizing natural language processing and pretrained transformers, it aims to classify videos as clickbait or non-clickbait effectively.

---

Workflow ⚙️

1. Fetching data using YouTube API:

   • Retrieved metadata from 750 videos, including comments, views, likes, thumbnails, and published dates.

2. Analyzing sentiments:

   • Used tools like Vader, TextBlob, and NLTK for sentiment analysis.
   • Implemented majority voting to determine the overall sentiment for each video.

3. Clickbait evaluation:

   • Combined sentiment analysis results and thumbnail information to declare whether a video is clickbait or not.

---

Progress 🚀

• Identifying sentiments using pretrained transformers like DeepSeek and Llama to enhance the analysis.

---

Learning Outcomes 🎓

1. Enhanced skills in using APIs for data extraction and handling large datasets effectively.
2. Gained practical experience in leveraging NLP techniques and transformers for sentiment analysis and classification.

---

Plugins and Libraries Used 🛠️

• ollama
• youtube-developer
• nltk
• vader
• TextBlob