Singapore Rental Hotspots

Analysis Date: 2026-02-20
Data Period: 2024-01 to 2025-01
Property Type: HDB rentals

Key Takeaways

The clearest finding

Rental premiums cluster geographically rather than appearing randomly. A small number of areas consistently command above-market rents, while coldspots show structural discounts.

What this means in practice

Investors can use hotspot status as a rental-demand signal, not as a guarantee of capital appreciation.
Homebuyers should avoid assuming that rental hotspots automatically justify an owner-occupier premium.
Value-seekers may find better space-for-budget trade-offs in persistent coldspots.

Core Findings

1. Hotspots are concentrated and statistically selective

Area	Gi* Statistic	Median Monthly Rent	Interpretation
Orchard	4.21	$3,200	Strong hotspot
Marina South	3.89	$3,450	Strong hotspot
Bukit Timah	3.45	$3,100	Strong hotspot
Woodlands	-3.21	$2,100	Strong coldspot
Yishun	-2.98	$2,050	Strong coldspot
Sembawang	-2.67	$2,000	Strong coldspot

Only 12 H3 cells qualified as 99% confidence hotspots, which is a useful reminder that true hotspots are limited rather than widespread.

2. The rent gap is large

Comparison	Value
Hotspot median rent	$3,200
Coldspot median rent	about 1,850 to 2,100 dollars
Typical gap	about $1,350 per month

Impact

For rental investors, location selection can dominate small property-level differences.
For owner-occupiers, that same premium should be tested against lifestyle benefit, not rental logic alone.

3. Cluster status has moderate persistence

Transition	Probability
Hotspot remains hotspot	58-62%
Coldspot remains coldspot	55-60%
Non-significant to hotspot	8-10%

Impact

Hotspot status is sticky enough to matter.
It is not so stable that it should be treated as permanent.

Decision Guide

For investors

Use hotspot status to screen for stronger rental demand.
Prefer established hotspots when stable income matters more than speculative upside.
Do not overpay for “emerging hotspot” narratives without evidence of improving fundamentals.

For buyers

Treat hotspot premiums as a tenant-market signal, not a universal valuation rule.
Coldspots can still be attractive when commute, amenities, and own-stay needs line up.

Technical Appendix

Data Used

Rental data: data/parquets/L1/housing_hdb_rental.parquet
Geocoding data: data/parquets/L2/housing_unique_searched.parquet (lat/lon)
Join key: block_street_name, filtered to successfully geocoded records (search_result == 0)
Date range: 2024-01 to 2025-01
Spatial aggregation: H3 resolution 8, median monthly_rent per cell
Coverage: 847 total cells, 623 with sufficient data for analysis

Methodology

Data join: L1 rental data joined to L2 geocoded properties via block+street_name key
Getis-Ord Gi* local statistic via esda.getisord.G_Local
Spatial weights: KNN (k=8) and Queen contiguity
Classification thresholds:
- Hotspot: z > 2.58 (99% confidence)
- Weak hotspot: z > 1.96 (95% confidence)
- Coldspot: z < -2.58 (99% confidence)
- Weak coldspot: z < -1.96 (95% confidence)
Permutations: 99 for significance testing
Transition probability analysis: persistence of hotspot/coldspot status across time periods

Technical Findings

12 H3 cells qualified as 99% confidence hotspots
Top hotspots: Orchard Gi* = 4.21, Marina South Gi* = 3.89, Bukit Timah Gi* = 3.45
Top coldspots: Woodlands Gi* = -3.21, Yishun Gi* = -2.98, Sembawang Gi* = -2.67
Rent gap: hotspot median $3,200 vs coldspot median ~$ 1,850- $2,100 ≈$ 1,350/month
Persistence rates: hotspot remains hotspot 58-62%, coldspot remains coldspot 55-60%
Transition rate: non-significant → hotspot only 8-10%, confirming true hotspots are limited

Conclusion

The Getis-Ord Gi* analysis identifies a small, statistically selective set of rental hotspots concentrated around central and city-fringe locations, with coldspots in northern suburbs. The ~$1,350/month rent gap between hotspots and coldspots is economically significant. Persistence is moderate (58-62%), meaning hotspot status is sticky but not permanent — investors should not treat it as a guaranteed perpetual premium. The low non-significant→hotspot transition rate (8-10%) confirms that emerging hotspot narratives require strong fundamental evidence. Key limitations: hotspot analysis summarizes spatial pricing only, not full asset returns; acquisition price may already capitalize the rent premium; areas can shift with new transport links, supply additions, or policy changes.

Scripts

scripts/analytics/analysis/spatial/analyze_spatial_hotspots.py — Getis-Ord Gi* with KNN and Queen weights