bellakhdarOmaima/Darija-NLP--Sentiment-Analysis-of-YouTube-Comments

# 📊 Project-NLP: Sentiment Analysis for YouTube Comments 🎥 ## General Description Project-NLP is an application designed for sentiment analysis based on YouTube comments. It enables real-time scraping of trending videos in Morocco or specific videos via their link, analyzes the sentiment of associated comments using multiple deep learning models (BERT, LSTM, CNN), and displays the results through an intuitive user interface. The collected data is stored in a MongoDB database, allowing further analysis of sentiments through a dashboard. The project stands out by utilizing a multilingual model (cardiffnlp/twitter-xlm-roberta-base-sentiment) for video link scraping and specific models (BERT, LSTM, CNN) trained on Darija and Arabic datasets for trending videos. It also includes data visualization features. ## 🌟 Features - 🔍 **Real-Time Scraping**: Scrape trending YouTube videos in Morocco or specific videos via their links. - 🧠 **Sentiment Analysis**: Classify YouTube comments into sentiments (Positive, Negative, Neutral) using state-of-the-art models like: - BERT - LSTM - CNN - CardiffNLP/twitter-xlm-roberta-base-sentiment for multilingual analysis. - 📊 **Data Visualization**: Generate insightful visualizations such as: - Word clouds for frequently used terms in comments. - Sentiment distribution charts. - 📁 **MongoDB Integration**: Store scraped data and predictions in a MongoDB database for future use. - 🖥️ **Interactive Dashboard**: View sentiment trends and analysis directly through an intuitive user interface. - 🌍 **Darija & Arabic Compatibility**: Models trained specifically for Moroccan Darija and Modern Standard Arabic datasets. - 🛠️ **Flask-Based Web Application**: Seamless integration with Flask for web-based interactions using HTML, CSS, and JavaScript. ## 📁 Project Structure ```bash Project-NLP/ ├── models/ # Pre-trained models for sentiment analysis │ ├── bert_classifier_model.pt │ ├── bilstm_cnn_model.pth │ ├── cnn_sentimen …