HOUSEPRICE API
Machine learning property valuation API & interactive Streamlit dashboard.
The Problem
Real estate platforms and home buyers need instant, reliable property valuation estimates via interactive interfaces and clean APIs without requiring heavy client-side computation or unvalidated inputs.
The Idea & Solution
A high-performance FastAPI microservice coupled with a reactive Streamlit UI that accepts property feature vectors (location, area, bedrooms, bathrooms, age) and outputs instant house price estimates using a trained machine learning model.
How I Built It & Architecture
Built with FastAPI and Pydantic for strict request/response data validation, paired with a Streamlit interface for live interactive predictions. Pre-trained Scikit-Learn regression models are loaded into memory on server startup for sub-10ms predictions.
[ User Streamlit UI / REST Client ]
│
▼
[ FastAPI + Pydantic Validator ]
(Schema & Type Verification)
│
▼
[ Feature Transformer Pipeline ]
(Encoding & Scaling)
│
▼
[ Scikit-Learn Valuation Model ]
│
▼
[ Predicted Price Response (<10ms) ] Tech Stack & Rationale
Key Retrospectives & What I Learned
Decoupling schema validation with Pydantic from model inference prevents invalid payloads from crashing the prediction pipeline and guarantees consistent JSON error contracts.