Skip to content

Latest commit

 

History

History
 
 

conversational-analytics

README.md

Welcome to the fullstack Conversational Analytics Generative AI application. This simple application lets customers send feedback while our system identifies Key Topics, Overall Sentiment, and Emerging Trends using a Large Language Model (LLM).

The application is still in development, but the basic features are ready for you to use, experiment and extend. Currently, it combines analytics for all users, but in the next update, I will separate the analytics for each user and add more important features related to conversational analytics.

Feel free to use this repo for your experiments and customize it to meet your needs.

🎥 YouTube

YouTube Video

🚀 Project Details

This project enables conversational analytics by analyzing customer feedback to uncover important topics, sentiment and emerging-trends. It leverages the power of Google's free Gemini-pro LLM API, providing valuable insights to help you understand your customers better. You can replace gemini-pro LLM API endpoint with any other LLM at main.py file.

Example:

Customer Feedback: I reached out for help with my account, and the support team was very responsive and helpful. I appreciate their quick assistance, but it would be helpful to have more self-service options available.

Conversational Analytics:

Key Topics: Account support

Overall Sentiments: Mixed (Positive | Negative)

Emerging Trends: Limited self-service options

🔥 Key Features

  • Real-time sentiment analysis
  • Topic detection and categorization
  • Trend identification and tracking
  • Role-based access control (Admin and User roles)
  • More...Work in Progress .........

🛠️ Technology | Tool Stack

  • Frontend: React
  • Backend: Python, FastAPI, uvicorn
  • Database: Mongodb (Uses inside docker so don't need to install separate at your system. Just use deafult URL given at the project)
  • AI Model: Google's Gemini-pro LLM API
  • Containerization: Docker (Need to install at your laptop/desktop)
  • Authentication: Custom JWT implementation
  • Postman: To create Admin user using rest api call with X-Admin-Key (SECRET_KEY). You can use CURL command as well if you have access to CURL at your system

📂 Project Structure

Conversational-analytics/
├── backend/
│ ├── test
│ ├── .gitignore
│ ├── auth.py
│ ├── Dockerfile
│ ├── main.py
│ └── requirements.txt
├── frontend/
│ ├── public/
│ └── src/
│ │ ├── components/
│ │ │ ├── AnalyticsDisplay.tsx
│ │ │ ├── FeedbackDisplay.tsx
│ │ │ ├── FeedbackForm.tsx
│ │ │ ├── Home.tsx
│ │ │ ├── Login.tsx
│ │ │ ├── Footer.tsx
│ │ │ ├── Header.tsx
│ │ │ ├── PrivateRoute.tsx
│ │ │ └── Register.tsx
│ │ ├── App.css
│ │ ├── App.tsx
│ │ ├── index.css
│ │ ├── index.tsx
│ │ ├── logo.svg
│ │ ├── react-app-env.d.ts
│ │ ├── reportWebVitals.ts
│ │ └── types.ts
│ ├── .gitignore
│ ├── Dockerfile
│ ├── package-lock.json
│ ├── package.json
│ ├── README.md
│ └── tsconfig.json
├── .env
├── .gitignore
├── Conversational-Analytics.postman_collection.json
├── docker-compose.yml
├── README.md
└── secret_key_generation.py

🚀 Getting Started

Prerequisites

Installation

  1. Clone the repository and open the folder at your IDE or terminal:

    git clone https://github.com/genieincodebottle/generative-ai.git
cd generative-ai/genai_usecases/conversational-analytics

  2. Set up your Google Gemini-pro API key:

    • Visit Google AI Studio

    • Create a new API key

    • Add the key to your .env file:

      GOOGLE_API_KEY=<your_api_key_here>

    • YouTube - How to setup Google's free Gemini Pro API Key

  3. Generate a secret key to create Admin User in other step:

    import secrets

secret_key = secrets.token_hex(32)  # Generates a 64 character hex string

print(secret_key)

    Copy the print output and add it to your .env file at following key:

    SECRET_KEY=<generated_secret_key>

  4. Build and run the Docker containers at your project folder:

    c:/<your_folder_location>/conversational-analytics> docker-compose up --build

    To close the application. Type Ctrl+C to stop the session or from Docker Desktop UI

    To remove the application from docker, run following command or use Docker Desktop UI

     c:/<your_folder_location>/conversational-analytics> docker-compose down

  5. Create an admin user using Postman (You can create Normal User using Register link at UI but to create Admin user, you need to run Rest API call using postman or using CURL command):

    • Import the Conversational-Analytics.postman_collection.json file at your local postman
    • After import, update the X-Admin-Key header with your SECRET_KEY
    • Send the request to create an admin user
    • Screenshots


    Note: If you want to use curl command instead of postman then run following curl command after changing X-Admin-key value and username & password

    curl --location 'http://localhost:8000/register' \
--header 'Content-Type: application/json' \
--header 'X-Admin-Key: <your secret-key generated in earlier step using python code>' \
--data '{
    "username": "suresh",
    "password": "suresh"
}'

  6. Access the application:

    • Frontend: http://localhost:3000/login
    • Backend API: http://localhost:8000

👥 User Roles

  1. Admin: Full access to analytics and feedback management
  2. User: Can submit feedback through the user interface

🔄 Development Workflow

  1. Make changes to the codebase as per your requirement like adding SECRET_KEY, GOOGLE_API_KEY
  2. Rebuild and restart containers: docker-compose up --build
  3. Test your changes

📸 Screenshots

  1. Login

  1. Register

  1. User Home Page

  1. Admin Home Page

  1. User Feedback

  1. Analytics

  1. Feedback Details

🛣️ Roadmap

  • Multi-user support with segregated feedback
  • Enhanced analytics capabilities
  • Streaming ......

Let me know if you face any issue running this application at your system.

Happy Coding! 🎉