Inspiration

Buying a home is one of the biggest financial decisions many Singaporeans will make. However, while a large amount of HDB resale data is publicly available, making sense of thousands of past transactions can be difficult, especially for first-time buyers. This inspired us to propose FlatFair, a data-driven tool that turns Singapore's open housing data into simple and useful insights for prospective HDB resale buyers. Rather than only showing past transactions, we want FlatFair to help users answer practical questions such as: What areas can I afford? Is the asking price of a flat reasonable compared with similar flats? How have prices in this area been changing?

What we aim to build

FlatFair will use Singapore's national open datasets, particularly HDB resale transaction data from data.gov.sg, to create an interactive housing affordability and price intelligence dashboard. Users will be able to select factors such as their budget, preferred town and flat type and explore:

Historical resale price trends and comparable transactions
Housing options that fall within their budget
An estimated price range based on similar flats and their characteristics
A short-term price outlook for selected towns and flat types The aim is not to predict the exact future price of a home, but to provide buyers with an additional data-driven reference point when comparing their options. ##How we plan to use Databricks We plan to build the project end-to-end on Databricks Free Edition. Public HDB datasets will first be ingested and cleaned before being stored as structured tables. We will engineer useful features such as remaining lease, price per square metre, town, flat type and transaction period. We then plan to experiment with machine-learning models for resale price estimation and time-series methods for short-term price trends. Model performance will be evaluated using historical data rather than simply presenting predictions without validation. The final results will be presented through an interactive dashboard designed for non-technical users. ##What we have learned so far While developing the idea, we realised that having access to open data does not necessarily mean that the information is easy to use. Raw transaction records can tell us what happened in the past, but additional processing and analysis are required to turn them into information that supports an actual housing decision. We also learned that housing prices are influenced by many interacting factors, such as location, flat type, floor area, storey and remaining lease. This means that simply comparing average prices may give users a misleading picture, motivating us to explore more detailed modelling. ##Challenges One major challenge will be ensuring that our predictions are useful without being misleading. Housing prices are affected by factors that may not be fully captured by the available datasets, so FlatFair should present predictions as estimates rather than guaranteed future prices. Another challenge will be balancing model complexity with interpretability. Our goal is not simply to build the most complicated model possible, but to produce results that an ordinary prospective buyer can understand and use. Ultimately, we hope FlatFair can make Singapore's public housing data more accessible and help prospective HDB buyers make more informed and transparent comparisons when considering their future home.