MakingNoMistakes

Inspiration

Buying a first HDB resale flat involves choices that a price chart alone cannot explain. A buyer might consider a lower storey, a smaller flat or a different town to stay within budget, but understanding what they would give up requires more than comparing two prices. FlatFair centres on a practical question: What would I trade off to stay within my budget? It brings historical transaction evidence, storey differences and financing scenarios into one comparison, helping first-time buyers understand their options and the assumptions behind them.

What FlatFair does

FlatFair is designed around a side-by-side comparison of two housing options. Buyers can compare flat type, floor area, remaining lease and historical resale transactions, then explore storey differences through an interactive 3D view. Both options use the same price and financing assumptions. Changing the price scenario, down-payment percentage, interest rate or loan tenure shows how each choice affects the estimated loan amount, monthly repayment and gap to the buyer’s budget. Upfront funding requirements remain visible separately. The 3D view gives storey comparisons a spatial explanation, while readable charts and comparison cards show the underlying evidence. Buyers can see where the options differ, which trade-offs reduce costs and whether their conclusions hold under a different scenario.

How we plan to build it

Databricks will power the data pipeline, combining HDB resale transactions from 2017 onwards, HDB Property Information, address coordinates and an official household-income benchmark. Raw data will be stored in Delta tables before cleaning and transformation. The pipeline will check missing values, unusual records, address matches and schema changes, then produce analysis tables for comparable transactions, price trends and affordability measures. The affordability index will express resale prices relative to annual benchmark household income, with the income definition and reference year clearly shown. This market-level measure will remain separate from the buyer’s own financing scenarios.

We also plan to evaluate six-month town-level price forecasts using chronological backtesting and comparisons against a simple baseline. MLflow will track experiments, while the interface will show forecast uncertainty and data cutoffs.

A React-based interface will combine the 3D explorer with comparison cards and charts, accessing prepared data through a backend and Databricks SQL. The initial two-week build will focus on one flat type, three towns and one neighbourhood for detailed 3D exploration. What we learned while developing the concept

A useful comparison depends on consistent assumptions. Applying different financing settings to each option can obscure the actual trade-off, which is why shared scenario controls are central to FlatFair. We also identified the importance of separating historical evidence, forecasts and hypothetical scenarios. Recorded transactions describe past sales; forecasts estimate future market trends; scenarios test assumptions. Keeping these distinct helps buyers understand what each result can reasonably tell them.

Another design lesson is that 3D should answer a meaningful question. In FlatFair, it helps explain storey differences, while the evidence and calculations remain accessible in a conventional comparison view.

Challenges

Finding genuinely comparable transactions is a key challenge. Flats differ in size, remaining lease, model and transaction date, and some blocks or storey bands have limited data. FlatFair will show comparison criteria and sample counts, and identify cases where the evidence is insufficient. The visualisation must also respect the limits of the data. Our 3D buildings will represent published storey bands rather than exact units, interiors or live listings. Price differences will not be presented as proof that storey height alone caused them.

Finally, affordability and forecasting require careful communication. Scenario calculations cannot establish loan eligibility or guarantee that a buyer can fund a purchase. Our goal is to make the trade-offs and uncertainty understandable, so buyers can ask better questions and make more informed comparisons.