YouTube Playlist Videos comments ETL and Sentiment Analysis

This project extracts video and comment data from a specified YouTube playlist and analyzes the sentiment of comments using a pre-trained RoBERTa model. The results are saved as CSV files for further analysis or visualization.

---

Features

• Extracts video details (title, views, likes, etc.) from a YouTube playlist.
• Scrapes comments for each video and cleans HTML content.
• Performs sentiment analysis on comments using the cardiffnlp/twitter-roberta-base-sentiment model.
• Saves the extracted data and analysis to CSV files.

---

Installation

1. Clone this repository or download the script.
2. Install the required libraries:

   pip install pandas numpy tqdm google-api-python-client beautifulsoup4 transformers torch matplotlib seaborn

3. Set up your YouTube Data API key by following the instructions here.

---

Usage

1. Set up API Key: Replace YOUR_API_KEY in the script with your valid YouTube Data API key.

2. Run the Script: Execute the Python script to extract data and analyze sentiments:

   python youtube_analysis.py

3. Output Files:

   • Video details are saved as linkin_park_youtube.csv.
   • Comment details with sentiment analysis are saved as linkin_park_youtube_comments.csv.

---

Code Details

1. Extract Video Information

• Retrieves video details such as:
  • Video ID
  • Title
  • Thumbnail URL
  • Published At
  • Channel ID
  • Channel Title
  • View Count
  • Like Count
  • Dislike Count (default is 0 if unavailable).
• Saves the data into a Pandas DataFrame.

2. Scrape Comments

• Extracts comments, commenter names, and publication times for each video in the playlist.
• Cleans HTML content in comments using BeautifulSoup.

3. Sentiment Analysis

• Analyzes the sentiment of comments using the cardiffnlp/twitter-roberta-base-sentiment model.
• Maps sentiment labels:
  • LABEL_2: Positive
  • LABEL_1: Neutral
  • LABEL_0: Negative
• Displays a progress bar for sentiment analysis using tqdm.

4. Save Results

• Saves both video and comment data with sentiment analysis to CSV files for further use.

---

Sample Output

Video Data

Video ID	Title	View Count	Like Count	Dislike Count
abc123	Video Title 1	5000	200	0

Comment Data with Sentiments

Video ID	Comment	Commenter	Sentiment
abc123	"Amazing content!"	User1	positive

---

Dependencies

• Python 3.8 or later
• Pandas
• NumPy
• Tqdm
• Google API Python Client
• BeautifulSoup4
• Transformers
• Torch
• Matplotlib
• Seaborn

---

Acknowledgements

• Pre-trained sentiment model: CardiffNLP's Twitter RoBERTa.
• YouTube Data API for playlist and comment extraction.

---

License

This project is licensed under the MIT License. See the LICENSE file for details.