Chat Support Sentiment Analysis

This project aims to analyze the sentiment of chat support conversations. By leveraging natural language processing (NLP) techniques, we can determine whether the sentiment of a conversation is positive, negative, or neutral.

Features

• Sentiment analysis of chat support conversations
• Visualization of sentiment trends over time
• Exportable reports for further analysis

Installation

1. Clone the repository:

   git clone https://github.com/sreshtha10/chat-support-sentiment-analysis.git

2. Navigate to the project directory:

   cd chat-support-sentiment-analysis

3. Create a data directory and download the dataset from Kaggle. Place the dataset inside the data directory.

   mkdir data
   # Download the dataset from Kaggle and place it inside the data directory

4. Install the required dependencies:

   pip install -r requirements.txt

5. Train the model by running model.py. The trained model will be saved in the trained_model directory.

   python model.py

6. Test the model by running predict.py.

   python predict.py

7. Run the Flask application by executing app.py.

   python app.py

Usage

The Flask service provided by this project is designed to integrate with the chat-support-service, a Spring Boot application available at Chat Support Service. Ensure both services are properly configured to work together for seamless sentiment analysis of chat conversations.

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

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