In the competitive Singapore real estate market, setting the right price for an HDB resale flat is critical. Real Estate agents need a reliable, data-driven tool to move beyond simple comparisons and provide clients with accurate, justifiable pricing advice.
The objective of this project is to analyze historical resale data to:
-
Identify the top factors that influence HDB resale prices.
-
Build a highly accurate predictive model to estimate a flat's value.
-
Deliver an interactive tool for agents to use to advise their clients.
In this project, I've developed an advanced XGBoost machine learning model that achieved an R² of 0.97, explaining 97% of HDB resale price variability in Singapore. This represents 49% reduction in average prediction error compared to a standard linear model, proving its ability to capture the complex dynamics of the market. These insights are delivered through a dynamic pricing tool for agents to use to advise their clients effectively.