kyleong / mindhive-subway

A web application filtering Subway outlets in Kuala Lumpur, Malaysia using Natural Language with AI. Built using React, FastAPI and LangChain with OpenAI.

Mindhive Subway

A web application designed to visualize Subway outlets in Kuala Lumpur, Malaysia. Built using React and FastAPI.

Features include an AI-powered query system that enables users to inquire about the locations and operating hours of Subway outlets.

Data retrieved from the Subway website, processed, and stored in a database.

Demo video

*Red pin - default, Blue pin - selected, Green pin - highlighted*

Please refer to the /client and /server directories for instructions on installation and execution

Tech stacks

Backend:

• Python (version 3.9.5)
• FastAPI - Backend framework
• SqlAlchemy - PostgreSQL ORM
• Selenium - Web Scraping
• Langchain Python SDK - SQL Database Agent (Postgres Dialect)
• Google Geocoding - Latitude Longitude API

Frontend:

• Typescript with Node JS (version 20.5.2)
• React - Frontend framework
• TailwindCSS - Styling
• Zustant - State managament
• React Google Maps @vis.gl/react-google-maps - Google Map

Deployment:

• AWS RDS - Postgres Database
• AWS EC2 - Code Runtime