Indian Residential Estate Price Simulator

Real estate prices depend on a mix of structural, locational, and categorical factors that don't always relate to price in a simple linear way. This project:

Simulates a realistic residential listings dataset for 6 Indian cities (Indore, Bhopal, Pune, Bangalore, Nagpur, Ujjain), deliberately including missing values, duplicate rows, and data-entry outliers.
Cleans and explores the data to understand what actually drives price.
Trains and compares two regression models — Linear Regression and Random Forest.
Wraps the better-performing model in an interactive Streamlit app so anyone can enter property details and get an instant price estimate.
Note on the data: the dataset used here is synthetically generated (see generate_data.py) rather than scraped, so the project can run fully offline and be reproduced by anyone without needing an API key or a scraped source. The generation logic still encodes realistic relationships (bigger area → higher price, farther from city center → lower price, etc.), so the cleaning, EDA, and modeling steps are all working with genuine signal and genuine noise — not toy numbers.