Conversational Data Retrieval on Real Estate Database

Table of Contents

1. Introduction
2. Installation
3. Usage
4. Results
5. Conclusion

Introduction

This project is an end-to-end conversational real estate database retrieval system using Streamlit UI, Google Palm LLM, and the Langchain framework. This system provides a user-friendly interface for querying real estate data conversationally. It interacts with a MySQL database for dynamic Q&A, utilizing few-shot learning and Hugging Face embeddings stored in ChromaDB to enhance query accuracy and relevance. Google Palm LLM handles natural language understanding, while Langchain manages the conversational flow.

Installation

To set up this project, follow these steps:

1. Clone the repository:

   git clone https://github.com/yourusername/conversational-real-estate-retrieval.git
   cd conversational-real-estate-retrieval
   

2. Install the required dependencies: use a virtual environment if required

   pip install -r requirements.txt

3. Set up the database: Ensure you have a MySQL database set up with your real estate data. You can import data using the sql file provided.

Usage

1. Start the Streamlit UI:

   streamlit run main.py

2. Enter your query in english

Results

The conversational system successfully retrieves and responds to queries about real estate data with high accuracy. Key features include:

• Natural language understanding powered by Google Palm LLM.
• Conversational management using Langchain.
• Enhanced query accuracy with few-shot learning.

Conclusion

This project demonstrates the integration of advanced AI technologies to create a dynamic and interactive real estate query system. The combination of Google Palm LLM, Langchain, and Hugging Face embeddings offers a robust solution for conversational data retrieval.