The project builds a full pipeline that scrapes real‑estate transaction data, trains an XGBoost model, and serves predictions via a FastAPI endpoint with SHAP explanations. It deploys on Azure using Functions, Container Apps, and static web hosting, providing a zero‑cost demo that auto‑re‑trains monthly. Designed for developers or data scientists who need a ready‑made, cloud‑native house‑price estimator, it offers more than a simple script by integrating CI/CD, monitoring, and a React front‑end. Compared to generic ML demos, it includes production‑grade deployment patterns and cost‑optimisation tricks.
View on GitHub →arthurhardman/previsor-imoveis