Our Methodology
How Propertylens Australia produces AI property price predictions, suburb analytics, and market data for 540,000+ Australian properties across every state and territory.
Data sources
Propertylens Australia aggregates multiple public and licensed data sources to build a comprehensive view of every indexed Australian property.
Public property sales records
Historical transaction records from state property registries. Includes sale price, date, type (auction/private), and days on market.
All states and territories
Active and recent property listings
Agent-listed properties from major real estate platforms. Includes asking prices, property descriptions, photos, and campaign data.
Refreshed continuously
Planning and zoning databases
Government planning overlay maps including flood zones, heritage overlays, bushfire areas, and zoning classifications.
Council-level data, updated quarterly
ABS Census demographics
Population, household income, age distribution, occupancy type, and household composition from the Australian Bureau of Statistics.
Suburb-level, ABS Census cycles
School catchment boundaries
Government and private school catchment zones linked to suburb profiles and property pages.
State education department data
Property attributes
Bedrooms, bathrooms, car spaces, land size, building size, year built, property type, and condition indicators extracted from listings and records.
540,000+ indexed properties
AI prediction methodology
A four-step pipeline that produces a price estimate, confidence score, and price range for any property in our database.
Comparable sales analysis
Identifies recent sales of similar properties within the suburb and nearby areas. Adjusts for differences in bedrooms, bathrooms, land size, building area, and condition. Sales within 6 months carry higher weight than older records. The radius expands automatically if insufficient comparables exist locally.
Feature-based valuation model
A regression model scores each property against suburb benchmarks: zoning classification, distance to CBD, school catchment quality, flood overlay status, and renovation indicators. Each feature is weighted by its statistical correlation with price in that specific suburb and property type.
AI analysis layer
An AI model analyses the property listing description, market sentiment, infrastructure context, and risk factors that numeric models may miss. It produces an independent price estimate with reasoning. This layer is particularly useful for unique properties where comparable sales are limited.
Adaptive blending and confidence scoring
The three predictions are blended using weights recalibrated weekly. If comparable sales have been more accurate than AI in a given suburb, their weight increases automatically. The final output includes a predicted price, a confidence-based price range, and a confidence score (low, medium, or high).
Quality and accuracy processes
How we ensure Propertylens Australia data and predictions are reliable, transparent, and continuously improving.
Public accuracy tracking
Every prediction is compared against actual sale prices. Results are published on the public accuracy dashboard — all predictions, not just successful ones.
View accuracy dashboardWeekly model recalibration
Prediction weights are adjusted weekly using verified sale outcomes. Suburbs where certain prediction layers perform better automatically receive higher weighting for that layer.
Outlier detection
Properties with unusual characteristics (very large land, mixed-use zoning, heritage constraints) are flagged with lower confidence scores to prevent misleading estimates.
Data currency monitoring
Suburb statistics are recalculated weekly. Stale data (suburbs with no recent sales) triggers lower confidence scores and methodology disclosures on the property page.
Limitations and disclaimers
Propertylens Australia estimates are automated and based on available data. They are not formal property valuations and should not replace professional advice from a licensed valuer, financial adviser, or legal professional.
Prediction accuracy varies by property type, location, and data availability. Properties with unique characteristics, limited comparable sales, or in areas with sparse data may receive lower confidence scores. The prediction engine discloses its confidence level on every estimate.
Data may contain errors, omissions, or delays. Property sales records depend on state registry processing times. Planning overlay data is sourced from government databases and may not reflect the most recent council decisions.
Propertylens Australia is a research tool. Any property purchase, sale, or investment decision should involve independent professional advice.
Frequently asked questions
Where does Propertylens Australia get its property data?
Propertylens Australia aggregates data from public property sales records across all Australian states and territories, active and historical listing data from real estate platforms, government planning and zoning databases, ABS Census demographic data, school catchment boundaries, and flood overlay maps. Data is refreshed continuously as new sales, listings, and government records become available.
How does the AI prediction engine work?
Propertylens Australia uses a three-layer prediction system: (1) comparable sales analysis finds recent similar sales and adjusts for property differences, (2) a feature-based regression model scores each property against suburb benchmarks, and (3) an AI layer analyses listing descriptions, market context, and risk factors. The three predictions are blended using adaptive weights recalibrated weekly based on actual sale results.
How accurate are Propertylens Australia predictions?
Every prediction is tracked against the actual sale price. Results are published on the public accuracy dashboard at app.propertylens.au/predictions, including all predictions — successful and unsuccessful. The system uses these results to continuously improve by adjusting prediction weights per suburb and property type.
How often is Propertylens Australia data updated?
Property sales data is updated as new records are published by state registries (typically within days of settlement). Active listing data is refreshed continuously. Suburb statistics (medians, growth rates, days on market) are recalculated weekly. Planning overlay and zoning data is updated quarterly or as government databases change.
Is a Propertylens Australia estimate a formal property valuation?
No. Propertylens Australia provides automated estimates and AI predictions for research and due diligence purposes. A formal property valuation requires a licensed valuer to physically inspect the property and issue a certified report. AI estimates are useful for shortlisting, investment screening, and understanding market context, but should not replace professional advice for lending or legal purposes.
Explore Propertylens Australia
House Price Prediction
AI-powered property valuations with confidence scoring
Accuracy Dashboard
Public tracking of every AI prediction vs actual sale
Property Archive
Browse 540,000+ Australian property addresses
Suburb Analytics
17,000+ suburb profiles with market data
Free Estimate
Instant property value estimates
Suburb Rankings
Top suburbs by growth, price, and affordability
Recently Sold
Latest property sales updated daily
Compare Suburbs
Side-by-side suburb comparison tool
Developer API
REST API for Australian property data