← Back to ML Projects
RealEstate AI – Property Valuation & Chatbot Platform
Predictive Modeling & Chatbot
Python
XGBoost
Claude API
Scrapy
FastAPI
React
PostgreSQL
Project Overview
RealEstate AI is an advanced, data-driven real estate platform that combines predictive modeling with conversational AI to provide instant, highly accurate property valuations and market insights.
Key Objectives & Methodology:
- Data Mining Pipeline: Developed an end-to-end web scraping infrastructure using Scrapy and Selenium to extract 10,000+ real estate listings from sources like OLX and Aqarmap, storing the structured data in PostgreSQL.
- Feature Engineering & Modeling: Created a robust data cleaning pipeline encompassing region normalization, outlier filtering, and target encoding. Trained a high-performing XGBoost pricing model achieving an R² above 0.85.
- AI Integration & Frontend: Built a Claude-powered chatbot backend with FastAPI, LangChain, and Redis. Paired it with a sleek React dashboard, enabling businesses to deploy branded AI assistants that deliver instantaneous property valuations and detailed market reports.
