Sentiment Analysis on Phone Call Transcripts

This project is a web-based application that performs sentiment analysis on uploaded call transcripts. Users can upload text files through a Streamlit-based UI, and the backend, powered by Flask, processes the files and returns sentiment analysis results.

The app uses a pre-trained Hugging Face Transformer model to classify sentiments into Positive, Negative, or Neutral, along with corresponding sentiment scores.

Project Structure

sentiment-analysis/
├── backend
│   └── app.py
|   └── demo.py    # to ananlyse the file locally on terminal
├── frontend
│   └── streamlit_app.py
├── uploads/
├── Sentiment_analysis_report.pdf
├── readme.md
├── requirements.txt
├── .gitignore
└── readme.md               

Getting Started

Prerequisites

• Python 3.11.11
• Required Python packages (listed in requirements.txt)

Installation

1. Clone the repository:

   git clone https://github.com/Princccee/Sentiment_analysis.git

2. Create a virtual environment:

   python3 -m venv .venv
   source .venv/bin/activate

3. Navigate to the project directory:

   cd sentiment-analysis

4. Install the required packages:

   pip install -r requirements.txt

5. Run the backend server:

   flask run  # spin up the server

6. Run the streamlit app in browser:

   streamlit run frontend/streamlit_app.py