HDB resale market intelligence & affordability forecasting

Inspiration

A first HDB flat is the largest financial commitment most young Singaporeans will make, and they make it with weak tools. Headline medians blur flat size, lease age and floor level. Resale prices rose 9.6% in 2024 by HDB's flash estimate, and the anxiety stuck even though the market has since cooled. HDB's Q3 2026 flash estimate shows resale prices easing for a third straight quarter. The question buyers face has shifted from "how high will prices go?" to "can I afford this, is it fair, and should I wait?" We wanted to answer that with public data and be open about what we can't predict.

What it does

FlatFair is a decision tool for first-time resale buyers. A household enters its income and gets:

A shortlist of towns and flat types it can afford under stated loan assumptions, ranked by headroom under the instalment cap. A fair-price band for comparable flats, with the mix-adjusted index shown beside the raw median so users can see where medians mislead. A six-month outlook by town with a prediction interval and a naive baseline for comparison.

How we built it

Data: HDB resale transactions (January 2017 onward), uploaded to a Unity Catalog Volume. Household income and mortgage inputs come from [sources] and are cited in the app. Pipeline: a Lakeflow pipeline with bronze, silver and gold layers. Expectations handle nulls, outliers and flat-type normalisation, and double as our data quality record. Delivery: an AI/BI dashboard with an income input.

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for HDB resale market intelligence & affordability forecasting

Built With

• databricks
• sql