Excel Plotter Chatbot

Project Description

The Excel Plotter Chatbot is an innovative application that combines natural language processing with data visualization. It allows users to interact with their Excel data through a chatbot interface, generating charts and graphs based on natural language queries. This project leverages advanced language models and plotting libraries to create a seamless experience for data analysis and visualization.

Images

Home page

Sample Query

Result

Features

• Natural language interface for querying Excel data
• Dynamic chart generation based on user queries
• Support for various chart types (line, bar, etc.)
• Real-time data processing and visualization
• Integration of LLM (Large Language Model) for query understanding
• Image verification to ensure chart accuracy

Models Used

1. BART (facebook/bart-large): Fine-tuned for understanding and processing natural language queries related to data visualization.
2. CLIP (openai/clip-vit-base-patch32): Used for image verification to ensure the generated charts match the user's query.

Workflow

Backend

1. train_model.py:

   • Script for fine-tuning the BART model on custom dataset and the model is saved as fine-tuned-bart-large

2. app.py:

   • Initializes Flask application
   • Defines API endpoints for handling requests
   • Manages static file serving

3. llm_agent.py:

   • Processes user queries using the fine-tuned BART model
   • Coordinates between data processing, chart generation, and image verification

4. data_processor.py:

   • Handles loading and preprocessing of Excel data

5. chart_generator.py:

   • Generates charts based on processed queries using Matplotlib

6. image_verifier.py:

   • Verifies generated charts against user queries using CLIP model

7. Now one can see the chart posted on the app

Frontend

1. index.js:

   • Entry point for React application

2. App.js:

   • Main component managing overall application state

3. Chatbot.js:

   • Handles user input and communication with backend API

4. ChartDisplay.js:

   • Renders the generated chart image

Usage Instructions

1. Enter your query in natural language (e.g., "plot the sales in the years with red line")
2. The system processes your query and generates a relevant chart
3. The chart is displayed on the screen

Project Setup

Cloning the Repository

git clone https://github.com/Transcendental-Programmer/LLM-Integrated-Excel-Plotter-App.git
cd LLM-Integrated-Excel-Plotter-App

Backend Setup

1. Create a virtual environment:

   python -m venv myenv
   myenv\Scripts\activate
   

2. Install required packages:

   pip install -r requirements.txt
   

3. Set up environment variables: Create a .env file in the backend directory and add the hugging face API key, you can get one for free from the web https://huggingface.co/.

4. change to backend

   cd backend
   

5. Run the train_model.py

   python train_model.py
   

6. Run the backend server:

   python app.py
   

The API should now be running on http://localhost:5000.

Frontend Setup

1. Navigate to the frontend directory:

   cd frontend
   

2. Install dependencies:

   npm install
   

3. Start the development server:

   npm start
   

Interact with the app opened on http://localhost:8080

Contributing

Contributions to improve the Excel Plotter Chatbot are welcome. Please feel free to submit pull requests or open issues to discuss proposed changes.

License

MIT License