Amenities and Property Prices in Singapore

Problem

Property-price research often focuses on a small number of obvious variables such as MRT access or school proximity.

That is useful, but it can also flatten the housing market into a handful of familiar signals. I wanted to test whether a broader amenity framework could explain price variation more effectively across Singapore’s property market.

Approach

This research combined long-run transaction data with a much wider amenity layer than is typical in smaller property studies.

• Property transaction data from 1990 to 2023 was used as the pricing backbone.
• Amenity information was assembled across dozens of categories, including transport, schools, malls, and other points of interest.
• SQL and Python were used to prepare, join, and structure the data.
• Both classical and machine-learning approaches were tested, including correlation work, LASSO-based modelling, Random Forest, and LSTM variants.

The aim was not just to produce a prediction score, but to build a broader framework for thinking about how neighbourhood context influences property values.

What the Research Found

The project indicated that including a wider amenity set improved model performance relative to more limited feature selections.

That matters because Singapore is a dense and highly differentiated urban system. A property’s surrounding context is rarely captured well by one or two convenience variables. Once the amenity layer becomes richer, the market picture becomes richer too.

Why It Mattered

This project sits closer to research than to dashboard work.

It helped shape how I think about housing-market analysis today: avoid over-relying on a narrow set of popular signals, and instead build models that reflect the actual complexity of how urban environments, access, and market value interact.