MAIGIC

MAIGIC: My AI Generates Insights and Charts

MAIGIC is a Retrieval Augmented Generation(RAG) based AI-powered tool designed to generate insightful analyses and visualizations from user's submitted CSV file. The project uses Langchain LLM and llama-3-70b-instruct to interpret datasets and create meaningful charts that help users understand underlying patterns and trends.

For deploying in local server(flask version) pull the entire repository and run app.py.

• LLM API creds would be made available on request

Features

• Data Insights: Takes user input data as a CSV file and requested query returns responses from a llama-3-70b-instruct hosted on IBM cloud using IBM WatsonX framework.
• Chart Generation: Internally utilises python's seaborm(matplotlib) and creates visualizations on the submitted data such as bar charts, line graphs, and more, creativity beyond limits.
• User Interface: Provides a visually appelaing & easy-to-use web interface for interaction.
• Pandas dataframe agent: Utilizes Pandas dataframe agent models to predict and analyze data.

Technologies Used

Deployments

There are two deployments of this project across two Python frameworks:

Flask Version

• Runs on a local server
• Supports robust designs and functionality using HTML, CSS, and JS
• Clone the repository: git clone https://github.com/lekh-ai/RAG.git

Streamlit Version

• Deployed on Hugging Face
• Streamlit Deployment Link

Both versions have different UI and frontend configurations. Unlike Streamlit/Gradio, Flask supports HTML, CSS, and JS for more robust designs and functionality.

UI Appearance ?️

Flask Version (localhost)

Streamlit Version (Hugging Face)

Example Test Run Case ?

the app uses a demo test studentdat.csv file for testing and generate queries and charts from it

Query Used: "Make a histogram for column Fjob"

API Response

Contact ? For more information or to get in touch, please email [email protected] Phone number : 9179041912