Machine Learning

Singapore HDB Resale Price Predictor

A gradient boosted regression model that predicts future resale prices of Singapore HDB flats from historical transaction data.

Overview

Singapore's HDB resale market publishes rich historical transaction data, but pricing a flat still comes down to gut feel for most buyers and sellers. This project trains a gradient boosted regressor on past resale transactions to predict future resale prices, using features like flat type, floor area, remaining lease, and location.

What it covers

  • Data cleaning and feature engineering on historical HDB resale transactions
  • Gradient boosted regression model training and hyperparameter tuning
  • Model evaluation against held-out transaction data

Tech stack

  • Python
  • Gradient boosted regression (scikit-learn / XGBoost-style modelling)
  • SQL for data wrangling
← Back to projects