# Open and Accessible Property Data for a Brisbane Development and Leasing Business

## Executive assessment

As at **14 September 2026**, a Brisbane-based property development and leasing company can reproduce a surprisingly large part of a PropTrack-style intelligence stack internally, but not all of it from open data alone.

The strongest strategy is an **“open core + licensed edge”** model:

**Open/public data should become the system of record for** addresses, parcels, zoning, planning overlays, school catchments, demographics, public transport, development applications, flood exposure, climate projections, crime context, pools, population forecasts and much of the suburb-level rental/development reporting. Queensland and Brisbane are unusually good environments for this because Queensland Government, Brisbane City Council, Translink, QPS, ABS and QGSO expose substantial machine-readable spatial and statistical data. citeturn15search0turn19search1turn19search7turn20search0turn20search5turn22search7

**Licensed or paid data is still required for the commercially valuable “last mile”:** comprehensive property-level transactions, ownership and titles at scale, current market-wide listings, rich property attributes, comparable sales, automated valuations, complete rental histories, listing photographs/text, portal search behaviour, buyer intent and market-wide agent/listing performance. Queensland's Valuer-General and Titles Queensland make sales and ownership information accessible, but property-level access is not an unrestricted free bulk feed. PropTrack and Domain explicitly commercialise these richer datasets through APIs and extracts. citeturn22search0turn22search1turn23search2turn23search3turn23search7

**The important architectural decision is therefore not “PropTrack or build it ourselves?”** It is to build the property intelligence platform yourselves and treat PropTrack, Domain, REA, QVAS and similar providers as replaceable data suppliers. That prevents a single vendor from becoming the company's database, analytics engine and user interface simultaneously.

My recommended target model is:

> **Government/open spatial data → internal property master → licensed transactions/listings → proprietary analytics → Power BI/BI dashboards, due-diligence reports and alerts.**

The resulting platform can own the valuable intellectual property—the property identity graph, spatial enrichment, scoring methodology, development pipeline, leasing intelligence, forecasts and portfolio analytics—while purchasing only the data that cannot economically or legally be recreated.

## The data-source landscape

The best foundation is a property master anchored on Queensland cadastral and address data.

Queensland's **Property Location Service** is particularly useful. It validates Queensland addresses and property descriptions, returns location information from a geocode, is free, requires no authentication for the public service and is updated daily. This is a very strong substitute for a commercial address-match endpoint within Queensland. citeturn19search1turn19search5

For a national address backbone, **G-NAF** remains the obvious choice. The August 2026 release is available through data.gov.au under an open licence, with geocoded addresses suitable for local indexing and fuzzy address search. citeturn19search2turn19search6

Queensland has also migrated its cadastre to the **Queensland Spatial Cadastral Fabric, or QSCF**. QSCF contains the spatial representation of current Queensland land parcels, legal Lot-on-Plan descriptions and associated attributes, and is updated weekly. Queensland also operates a time-aware service incorporating DCDB history from 1997 to April 2026 and QSCF parcels thereafter. That creates a powerful open backbone for parcel identity and parcel change detection. citeturn19search7turn19search8turn19search0

Brisbane City Council adds a second rich layer. Its open-data platform supplies spatial/CSV datasets and APIs, including City Plan zoning and flood metrics, while **Development.i** exposes current and historical development applications, application documents, map searches and notifications. citeturn15search0turn23search1turn15search1

At the other end of the spectrum sit PropTrack and Domain. PropTrack currently advertises APIs for transactions, address suggest/match, property histories and attributes, suburb market data, and listings search/details. Domain's developer platform exposes packages covering address suggestions, agents/listings, properties/locations, price estimation, property enrichment, rental AVMs and schools; it says new developer projects receive immediate access to its Agencies & Listings and Properties & Locations packages, subject to the applicable API terms. citeturn23search3turn23search2turn23search6

REA's Partner Platform is narrower and strategically different. It is particularly useful for a company's **own authorised campaigns and listings**: the Campaign API exposes listing/project advertising performance and engagement metrics, while the Leads API supplies consumer enquiries and associated contact/listing metadata. citeturn17search1turn17search2turn17search5

The practical source hierarchy should therefore be:

| Tier | Role in the platform | Primary sources |
|---|---|---|
| **Open authoritative** | Default source wherever possible | QLD Open Data/QSpatial, Brisbane City Council, ABS, QGSO, Translink, QPS, RTA, Department of Education, CER |
| **Public lookup / government paid** | Legal identity, ownership and official transaction verification | Titles Queensland, Valuer-General, Office of Fair Trading |
| **Commercial API / extract** | Listings, rich attributes, transactions at scale, AVMs, portal behaviour | PropTrack, Domain, QVAS brokers, REA Partner Platform |
| **First-party** | Highest-value proprietary behavioural data | CRM, leasing system, development pipeline, website analytics, REA leads/campaigns |
| **Derived internally** | Competitive intellectual property | Liveability, yields, risk scoring, forecasts, development scores, propensity models |

That hierarchy avoids paying commercial licence fees for information government already publishes while preserving access to datasets government does not provide.

## Coverage of the requested PropTrack-style data points

The following matrix maps every requested capability to the most realistic source and indicates whether I would **build**, **buy**, or combine the two.

| Data point | Best source and accessibility | Recommended approach |
|---|---|---|
| **Agency Lookup** | Queensland Office of Fair Trading maintains a free public register covering licensed property agents, salespeople and corporations. citeturn23search0turn23search12 | **Build/public lookup.** Supplement corporate entities with free ABN Lookup web services. citeturn18search2turn18search10 |
| **Agent Lookup** | OFT register supports person/company/licence searches and is updated overnight. citeturn23search12 | **Build lookup**, but don't bulk-harvest it for marketing: the register's conditions expressly restrict bulk requests for marketing purposes. citeturn23search8 |
| **Agent Details** | OFT records can include licence holder, registered business name/place of business, status, licence number and class; formal extracts are paid. citeturn23search0 | **Partial open.** Enrich business identity through ABN Lookup; buy Domain/PropTrack data for richer commercial agent information. citeturn18search2turn23search2 |
| **Points of Interest** | Brisbane Council open datasets plus OpenStreetMap for broad POI coverage. OSM data is available under ODbL conditions. citeturn15search0turn13search3 | **Build.** Load POIs into PostGIS and calculate walking/driving distances. |
| **Schools** | Queensland Department of Education publishes state-school site locations; non-state information can be supplemented through education directories or commercial APIs. citeturn20search0 | **Build core; commercial optional.** |
| **School Zoning** | Department of Education publishes 2026 primary, junior-secondary and senior-secondary catchment boundaries as KML, updated annually. citeturn20search0turn20search8 | **Build.** Spatially intersect each property point/parcel with catchments. |
| **Liveability** | No single authoritative open “liveability” truth is required: ABS demographics/SEIFA, schools, transport, parks/POIs, crime and environmental data provide the underlying components. citeturn20search2turn20search10turn20search5turn21search7 | **Build proprietary index.** This is an opportunity to create IP rather than buy a black-box score. |
| **Demographic** | ABS Census DataPacks provide data down to SA1; QGSO supplies more recent population statistics and projections. citeturn20search2turn20search6turn20search11 | **Build.** Use SA1 for neighbourhood characteristics and SA2 for more current market/population measures. |
| **Hospital Zoning** | Queensland publishes Hospital and Health Service boundaries, but these are health-service administrative boundaries rather than precise single-hospital residential catchments. citeturn2search2 | **Approximate/build.** Model HHS membership plus nearest hospital/travel time. Do not label it an exact hospital catchment unless an authoritative catchment exists. |
| **Transport Proximity** | Translink publishes GTFS schedules for SEQ bus, rail, ferry and tram plus GTFS-Realtime, with no authentication and CC BY licensing. citeturn20search5turn20search1 | **Build.** Calculate distance to stops, service frequency, peak frequency and journey time to CBD/employment centres. |
| **Property Zoning** | Brisbane City Plan zoning is directly available as open spatial data/API and is regularly updated. citeturn23search1turn23search5 | **Build. Essential.** Also ingest relevant City Plan overlays rather than storing only the headline zone. |
| **Transacted Sales** | Valuer-General sells property sales information; QVAS brokers supply valuation/sales data commercially. QGSO publishes aggregated dwelling-sales statistics. citeturn22search0turn22search7 | **Buy for parcel-level scale; open for aggregated reporting.** |
| **Ownership Details** | Titles Queensland current title searches identify current owner and registered interests including mortgages, easements, covenants, leases and caveats. citeturn22search1 | **Paid government source.** Do not build an ownership database by scraping public interfaces. |
| **Title References** | Titles Queensland title search and QLD cadastral/property identifiers support the link from parcel to title. citeturn22search1turn19search7 | **Store in master property model** where lawfully obtained. |
| **Residential Buy & Rent Listings** | PropTrack Listings API and Domain Agents & Listings. REA's Listing Export API is available for authorised customer data rather than as a free universal listings dump. citeturn23search3turn23search6turn10search6 | **Buy/license.** Current listing inventory is one of the major gaps in an open-only stack. |
| **Commercial Buy & Lease Listings** | PropTrack covers commercial data; REA Partner APIs support realcommercial.com.au enquiries/performance, and licensed commercial feeds are the appropriate source. citeturn23search7turn17search2 | **Buy/license.** |
| **Crime** | QPS publishes downloadable reported-offence trends and an online crime map. The map receives QPRIME extracts daily with about a one-day lag, but displayed locations are deliberately generalised rather than exact offence locations. citeturn21search3turn21search7 | **Build risk indicators**, preferably from official division/LGA rates plus local density rather than presenting map points as exact address crimes. |
| **Climate Projections** | Queensland Future Climate provides CMIP6-based high-resolution projections, dashboards and downloadable information for heat, precipitation, drought, wetness and fire-weather variables through the century. citeturn21search2turn21search6 | **Build climate exposure layer.** |
| **Environmental Hazards** | Brisbane Flood Awareness, parcel-level Flood Awareness metrics and Queensland flood/hazard layers provide open inputs. Council's parcel metrics are refreshed roughly four-weekly but expressly are not a substitute for conveyancing/due diligence. citeturn21search1turn21search9turn21search0 | **Build screening system; retain statutory reports/professional review for actual acquisitions.** |
| **Natural Hazard Ratings** | Open sources provide hazard inputs, not one universally authoritative parcel-level risk score. Brisbane flood, statewide flood, climate and other hazard layers can be combined. citeturn21search0turn21search9turn21search2 | **Build a transparent internal rating.** Store the underlying hazards separately from the derived score. |
| **Forecasts** | QGSO provides current population and housing/development series; PropTrack provides HPI data and commercial market datasets. citeturn20search11turn22search7turn17search0 | **Build models.** Stronger suburb-level property-price forecasting requires licensed transaction/listing history. |
| **Suburb Price Growth** | QGSO Housing Profiles include residential dwelling sales at SA2 scale; PropTrack publishes HPI data and market growth data. citeturn22search3turn17search0 | **Build for internal trends; buy if you need a hedonic index comparable to PropTrack.** |
| **Suburb Rental Yields** | RTA publishes quarterly median rents based on new bond lodgements by suburb/postcode/LGA; QGSO publishes residential dwelling sales. citeturn22search2turn22search3 | **Build indicative gross yields** by combining rent and price data with compatible property types/geographies. |
| **Home Price Index** | PropTrack publicly exposes the Home Price Index page and latest downloadable data; the index is calculated frequently and reported monthly. citeturn17search0 | **Use directly as an external benchmark**, not as your only internal price series. |
| **Valuation Insights** | Queensland statutory land valuations are available, while PropTrack and Domain supply commercial price-estimation/AVM products. citeturn22search0turn23search2turn23search3 | **Hybrid.** Land value ≠ improved-property market value; use AVM/licensed sales for market valuation. |
| **Address Suggest** | G-NAF can be indexed locally; Domain and PropTrack offer explicit address-suggestion APIs. citeturn19search2turn23search2turn23search3 | **Build cheaply** using G-NAF/Postgres search, with commercial API as fallback. |
| **Address Match** | Queensland Property Location Service validates addresses/property descriptions and returns geocoded location data free without authentication. citeturn19search1 | **Build around PLS.** Add G-NAF for national coverage and fuzzy matching. |
| **REA Site Engagement** | REA Campaign API gives authenticated customers/partners campaign performance including views and engagement events for authorised listings. citeturn17search1turn17search5 | **Integrate REA**, not open data. |
| **REA Search Insights** | Broader portal search/consumer behaviour is proprietary REA/PropTrack information; PropTrack explicitly markets unique consumer-behaviour signals. citeturn23search7 | **Buy where commercially justified.** Build first-party search analytics for your own website separately. |
| **Potential Buyers** | Consumer enquiries are available through REA Leads API for participating customers; broader buyer-intent datasets are proprietary. citeturn17search2turn23search7 | **Own first-party buyer graph + licensed REA leads.** |
| **Modelling** | PropTrack sells proprietary analytics and valuation models; open datasets can supply much of the model feature set. citeturn23search7turn23search11 | **Build models where differentiating; buy commodity AVMs.** |
| **Listing Engagement** | REA Campaign API includes portal metrics such as page views, enquiries, phone reveals and photo interactions. citeturn17search1turn17search5 | **Direct REA integration.** |
| **Agent Performance** | REA campaign/leads data plus internal CRM can produce performance metrics for participating agents/listings. citeturn17search1turn17search2 | **Build first-party performance reporting.** Market-wide competitor performance remains commercial. |
| **Predicted to List** | No comparable authoritative open prediction feed was identified. PropTrack markets predictive analytics/unique behaviour signals. citeturn23search7turn23search11 | **Build only after acquiring licensed listing/transaction history and sufficient first-party data.** |
| **Predicted to Sell** | Same data problem: good prediction requires historical listing, price, sale, property and market variables. PropTrack supplies these commercially. citeturn23search3turn23search7 | **Build later.** |
| **Predicted to Re-finance** | Credit-related personal information is subject to Australia's specialised Privacy Act credit-reporting regime; open property datasets do not expose mortgage balances, rates or refinance intent. citeturn18search1turn18search0 | **Do not treat this as an open-data feature.** Restrict to appropriately consented/authorised first-party financial data and legal/privacy review. |
| **Listing Alerts** | Domain provides webhooks and listing APIs; PropTrack offers listing search. citeturn23search6turn23search3 | **Buy feed, build alert engine.** |
| **Sale Alerts** | Licensed transactions/listing-status feeds provide timelier signals than one-off government sale searches. citeturn22search0turn23search3 | **Buy feed, build matching/alert layer.** |
| **Properties For Sale** | Current inventory from PropTrack/Domain listing feeds. citeturn23search3turn23search6 | **Commercial.** |
| **Properties Sold** | Official sales via Valuer-General/QVAS; commercial listing/transaction APIs improve timeliness and accessibility. citeturn22search0turn23search3 | **Hybrid.** |
| **Properties For Rent** | Current listing feeds through PropTrack/Domain; RTA is historical/aggregate rent rather than current listing inventory. citeturn23search3turn22search2 | **Commercial feed.** |
| **Properties Leased** | RTA bond lodgements provide aggregate evidence of newly established rentals, not a comprehensive property-by-property “leased listings” feed. citeturn22search2 | **CRM/portal feed for exact leased properties; RTA for market benchmarking.** |
| **Developments** | Brisbane Development.i provides current/historical DAs; QGSO publishes quarterly development pipeline indicators including multiple-dwelling MCU stock, lot registrations and approvals. citeturn15search1turn15search11turn22search7 | **Build. This can become an excellent proprietary competitor/development pipeline.** |
| **Advised Sales Results** | Domain provides auction/property datasets and PropTrack commercial transaction/listing data; official settled transaction records come later. citeturn23search2turn23search15turn22search0 | **Commercial for immediate/advised results; official transaction feed for confirmation.** |
| **Property Attributes** | Open cadastre/address/building datasets give structural/spatial attributes, while PropTrack/Domain supply richer dwelling attributes. citeturn19search7turn23search3turn23search2 | **Hybrid.** |
| **Property Boundaries** | QSCF is the open spatial representation of every current parcel and Lot-on-Plan in Queensland, updated weekly. It should not be confused with a legal survey of a boundary. citeturn19search7turn19search0 | **Build from QSCF.** |
| **Rental History** | RTA's public data is aggregate; property-level rental/listing history is a commercial data asset. citeturn22search2turn23search7 | **Licensed data + own tenancy/lease records.** |
| **Sale History** | Valuer-General property-sale searches and licensed QVAS/PropTrack/Domain data. citeturn22search0turn23search3 | **Buy/ingest.** |
| **Property Details** | Address, lot/plan, geometry, zoning, overlays and location can be built openly; bedrooms, bathrooms, car spaces and marketing attributes generally require listings/property feeds. citeturn19search1turn19search7turn23search1turn23search3 | **Hybrid master record.** |
| **Estimated Property Valuation** | Domain Price Estimation and PropTrack valuation products provide ready-made estimates. citeturn23search2turn23search11 | **Buy initially.** Only build an AVM once enough clean transaction/property history exists to justify model governance. |
| **Estimated Rental Yield** | Domain offers Rental AVM; internally, RTA median rents can be combined with sales/valuation data. citeturn23search2turn22search2 | **Build suburb/portfolio yield; buy property-specific rental AVM where required.** |
| **Transaction History** | Valuer-General/QVAS plus licensed commercial transaction feeds. citeturn22search0turn23search3 | **Licensed ingestion.** |
| **Comparable Sales** | Requires a suitable transaction universe plus property attributes; PropTrack and Domain package these commercially. citeturn23search3turn23search2 | **Build comparable-selection logic on licensed transactions.** Owning the comp algorithm is more strategic than owning the raw sale feed. |
| **Images / Video** | PropTrack explicitly markets property images associated with its listing/property data; portal imagery is not a general-purpose open asset. citeturn23search7 | **Own media + licensed portal content.** Do not architect the system around scraping portal photographs. |
| **Property Text** | Current descriptions come from authorised listing feeds/own CRM. Domain and PropTrack supply listings APIs. citeturn23search3turn23search6 | **Own/licensed.** |
| **Features — pools** | Queensland's open Pool Safety Register contains property addresses and pool information, though the bulk dataset visible on the portal is currently dated August 2024; QBCC's live register remains searchable in 2026. citeturn16search1turn16search5turn16search2 | **Useful open enrichment, but monitor staleness.** |
| **Features — solar** | Clean Energy Regulator publishes solar installation counts/capacity monthly by postcode, current to July 2026 in the latest release—not by individual property. citeturn16search0turn16search4 | **Use for neighbourhood solar penetration**, not parcel-level solar identification. Property-level feature needs listings/imagery/consented data. |

## Where open data is exceptionally strong

Brisbane's **planning and development intelligence** is the clearest opportunity to outperform a generic national property product.

Development.i currently exposes applications by address, suburb, application number and map, including recently received/decided applications and their documents. QGSO's September 2026 residential-development release adds quarterly series for broadhectare supply, approvals, certifications, lot registrations, sales, dwelling approvals and multiple-dwelling MCU pipelines. For example, the current QGSO release reports **52,971 potential new multiple dwellings in Brisbane's closing stock of MCU approvals at the end of the March 2026 quarter**. citeturn15search1turn22search7

Those sources could power an internal **development pipeline intelligence product** showing every site in the portfolio alongside neighbouring applications, approved apartment supply, lot production, competing projects and planning constraints. That is more useful to a developer than a generic property-data page.

Brisbane's flood data is similarly rich. Council publishes a parcel-metrics table that already contains the results of spatial intersections between property lots and flood layers. As of June 2026 it was described as a roughly four-weekly dataset. Council explicitly warns it is an awareness product rather than a substitute for formal due-diligence/planning information; that distinction should be preserved in any internal report. citeturn21search9turn21search1

Transport accessibility can be modelled with far greater sophistication than simply “500 metres from train station.” Translink's open GTFS provides routes, stops, timetables and service patterns across SEQ and its real-time feed supplies live vehicle/trip information without authentication. citeturn20search5turn20search1

An internal metric could therefore distinguish:

> 250 m from a bus stop with six weekday services  
> **versus**  
> 450 m from a station with 10-minute peak frequency and a 17-minute CBD journey.

Both are “near transport”; commercially, they are not remotely equivalent.

School zoning can likewise be deterministic instead of approximate. Queensland's 2026 catchment polygons cover primary, junior secondary and senior secondary catchments, enabling a simple point-in-polygon calculation whenever a property is ingested. citeturn20search0

Demographics should combine slow-moving and fast-moving sources. ABS Census DataPacks provide rich SA1-level household, family and dwelling characteristics, while QGSO's current population projections extend to 2046. Queensland's 2025-edition projections, released in 2026, project Brisbane LGA's medium-series population from about **1.263 million in 2021 to about 1.669 million in 2046**. citeturn20search2turn20search7

That lets a developer report not merely “demographics today” but **demographic trajectory**, which is considerably more useful for acquisition decisions.

Rental analytics are another open-data sweet spot. The RTA's quarterly data is based on newly lodged rental bonds and reports median rents by dwelling type and geography. QGSO Housing Profiles currently include June-quarter 2026 median rent, March-quarter 2026 dwelling sales, May 2026 building data and June-quarter 2026 residential lot registrations down to SA2. citeturn22search2turn22search3

The combination is sufficient to build credible suburb-level measures such as:

\[
\text{Indicative gross yield}
=
\frac{52 \times \text{median weekly rent}}{\text{median dwelling price}}
\]

The metric should be labelled **indicative**, because the rent and sale samples are not necessarily the same properties.

## What should be bought rather than rebuilt

Three areas are poor candidates for an open-data-only strategy.

**Property transactions and ownership** are the first. Queensland's Valuer-General currently charges **$10.54 GST-free for an abbreviated sale** and **$21.13 GST-free for a full sale search**, both of which include vendor/purchaser information; QVAS data is also distributed through participating brokers. Titles Queensland separately supplies current title and historical-title products. citeturn22search0turn22search1

That is fine for occasional due diligence but inefficient as the backbone of a business intelligence warehouse. At development-company scale, I would seek **a licensed QVAS/transaction feed or PropTrack/Domain transaction licence**, then use official title searches selectively for legal verification.

**Current listings** are the second. Government has no equivalent of a continuously refreshed, comprehensive inventory of residential and commercial properties currently advertised for sale or lease. PropTrack exposes listing search/detail APIs, while Domain exposes agents/listings and related property packages. citeturn23search3turn23search6

The better commercial question is therefore not “Can we avoid paying for listing data?” but:

> **What is the lowest-cost licensed feed that gives us the rights, refresh frequency, history and fields our own analytics actually require?**

Buying raw listing access but building your own aggregation, competitor analysis, supply pipeline, time-on-market and rent-change analytics preserves most of the strategic value internally.

**AVMs and rich property attributes** are the third. Building a serious automated valuation model requires a deep history of clean transactions, comparable-property attributes, temporal effects, location variables and model monitoring. PropTrack states that its AVM covers more than 12.6 million residential properties, while Domain offers both Price Estimation and Rental AVM packages. citeturn23search11turn23search2

For a property developer rather than a property-data company, I would initially **buy AVMs and build decision models around them**. An internal model answering “Should we acquire/develop/lease this?” is strategically valuable; recreating a national AVM primarily to save an API fee usually is not.

REA's proprietary behavioural layer is also fundamentally non-replicable with government data. Campaign APIs expose engagement with your authorised campaigns and the Leads API delivers consumer enquiries; PropTrack separately markets consumer-behaviour signals derived from realestate.com.au. citeturn17search1turn17search2turn23search7

The company should ingest these signals into its own warehouse so that, over time, **REA becomes an input rather than the place where historical campaign intelligence lives**.

## Recommended in-house property intelligence platform

The centrepiece should be a **canonical property master**, not a collection of disconnected dashboards.

A practical schema would create one internal `property_id` and map every known external identifier to it:

```text
PROPERTY_MASTER
│
├── internal_property_id
├── canonical_address
├── G-NAF address identifier
├── Queensland address/location identifier
├── latitude / longitude
├── lot / plan
├── cadastral parcel identifier
├── title reference        [licensed/restricted]
├── valuation identifier  [where licensed]
├── LGA / suburb / postcode
├── SA1 / SA2 / SA3
└── geometry
```

Queensland's free Property Location Service, current Queensland address data, G-NAF and QSCF supply nearly everything necessary to establish that identity layer. citeturn19search1turn19search9turn19search2turn19search7

Every other dataset should attach to that property master through either a **key join** or a **spatial join**:

```text
                         ┌── zoning / overlays
                         ├── flood / hazard
                         ├── school catchments
ADDRESS → PARCEL → PROPERTY ── transport accessibility
                         ├── demographics / SEIFA
                         ├── development applications
                         ├── transactions
                         ├── listings
                         ├── rents
                         ├── ownership / title
                         └── CRM / enquiries / campaigns
```

A Brisbane-focused organisation does not initially need a giant enterprise data stack. A technically competent team could achieve this with:

**PostgreSQL + PostGIS** as the operational spatial warehouse, object storage with Parquet files for raw histories, a transformation layer such as SQL/dbt, scheduled API/file ingestion, and Power BI or another preferred BI platform for presentation. The data model matters vastly more than which cloud logo sits on top of it.

I would structure the data into three layers:

| Layer | Purpose | Example |
|---|---|---|
| **Raw** | Immutable copy of source data | Original QSCF, GTFS, ABS CSV, RTA workbook, listing API payload |
| **Conformed** | Cleaned and attached to common geography/property IDs | `property_zoning`, `property_school_zone`, `property_transactions` |
| **Analytics** | Business-ready facts and scores | `development_score`, `gross_yield`, `market_velocity`, `risk_score` |

Every record should carry at least:

```text
source
source_record_id
source_effective_date
ingested_at
licence_class
confidence
property_match_method
property_match_score
```

That seemingly boring metadata becomes critical. A flood polygon, a portal listing, a title record and an inferred liveability score have radically different legal and analytical meanings.

The ingestion cadence can be source-specific rather than attempting artificial real-time updates:

| Dataset family | Sensible refresh |
|---|---|
| Queensland address validation | Daily |
| QSCF cadastre | Weekly; official dataset is updated weekly. citeturn19search7 |
| Brisbane City Plan | Change-driven / weekly monitoring. citeturn23search1 |
| Development.i | Daily |
| Listings | Near real-time / hourly depending licence |
| REA enquiries | Webhook/near real-time. citeturn17search2 |
| Translink static GTFS | On publication; real-time only where operationally useful. citeturn20search5 |
| Crime | Monthly for stable reporting; QPS map source can be one day behind for exploratory analysis. citeturn21search3turn21search7 |
| RTA rent | Quarterly. citeturn22search2 |
| QGSO development | Quarterly. citeturn22search7 |
| ABS Census | Census cycle |
| QGSO population | Release-driven |
| Climate projections | Release-driven |
| School catchments | Annual. citeturn20search0 |

## Analytics worth owning internally

The strongest business case is not merely to copy PropTrack's fields. It is to use the raw inputs to create measures specifically relevant to development and leasing.

**A site-acquisition score** could combine zoning/development potential, lot size/configuration, flood and environmental constraints, proximity to high-frequency public transport, school catchments, current rents, sales, competing development approvals and demographic growth. All but the transaction/listing component can largely be sourced from open Queensland/Brisbane data. citeturn23search1turn21search9turn20search5turn20search0turn22search3turn22search7

A conceptual model might be:

\[
S_{\text{development}}
=
w_1 Z
+
w_2 T
+
w_3 G
+
w_4 R
-
w_5 H
-
w_6 C
\]

where:

- \(Z\) = planning/zoning potential,
- \(T\) = transport accessibility,
- \(G\) = demographic/population growth,
- \(R\) = rental/market attractiveness,
- \(H\) = hazard constraints,
- \(C\) = competing approved development.

The important part is not the exact weights. The intellectual property is the **transparent methodology and historical validation against actual acquisition/development performance**.

A **leasing intelligence score** can similarly combine RTA bond rents, active listings, time on market, supply changes, first-party enquiry levels and REA campaign conversion data. RTA provides the market benchmark while authorised REA APIs provide direct campaign and enquiry signals. citeturn22search2turn17search1turn17search2

A **proprietary liveability/accessibility model** should use actual travel utility rather than arbitrary circles around POIs. The inputs are available: Translink supplies service patterns, Education publishes school sites/catchments, ABS supplies socio-demographics and Brisbane/OSM supplies amenity locations. citeturn20search5turn20search0turn20search2

For example:

```text
Accessibility score
 = CBD peak travel-time score
 + major employment-centre access
 + high-frequency transit access
 + school accessibility
 + retail/health/park accessibility
```

This creates something more defensible than “within 1 km of a train station”.

A **natural-hazard score** should remain decomposable. Instead of storing:

```text
Risk = 7.3 / 10
```

store:

```text
river_flood_exposure
creek_flood_exposure
overland_flow_exposure
storm_tide_exposure
bushfire_exposure
landslide_exposure
heat_projection
extreme_rain_projection
```

and then calculate the composite score. Brisbane Council itself emphasises that its flood awareness information is general and that formal planning information may be required for due diligence. citeturn21search1turn21search9

That design means an acquisition committee can understand **why** a site received its risk score.

Predicted-to-list and predicted-to-sell models are feasible later once several years of licensed listing/sale history have accumulated. Useful features would include time since acquisition, prior listing history, local turnover, property characteristics, development applications, neighbourhood supply changes and market momentum. PropTrack's commercial offering illustrates that these predictions sit on top of deep transaction, listing and behavioural datasets rather than on basic open cadastral data alone. citeturn23search3turn23search7

**Predicted-to-refinance is different.** Consumer credit information is regulated under Part IIIA of the Privacy Act and associated credit-reporting rules, while the Australian Privacy Principles govern covered organisations' handling of personal information more broadly. A development/leasing company should not casually infer or target individual owners' refinancing behaviour from combined personal/property datasets without a clearly lawful data basis and privacy review. citeturn18search0turn18search1

## Recommended delivery roadmap and commercial strategy

The first release should deliberately avoid expensive data licences. Build the **property and location intelligence foundation** first:

| Initial capability | Source |
|---|---|
| Address matching | Queensland Property Location Service + G-NAF |
| Parcel identity/boundaries | QSCF |
| Zoning and planning overlays | Brisbane City Plan |
| Development applications | Development.i |
| Flood | Brisbane Flood Awareness parcel metrics |
| School zones | Queensland Education |
| Public transport | Translink GTFS |
| Demographics | ABS + SEIFA |
| Population growth | QGSO |
| Rental benchmarks | RTA |
| Development supply | QGSO |
| Crime | QPS |
| Climate projections | Queensland Future Climate |
| Agent licence validation | OFT |
| Business entity validation | ABN Lookup |
| Pool feature | QBCC register |

These sources are either open or publicly accessible and collectively establish the difficult part of the platform: a **normalised spatial property graph**. citeturn19search1turn19search7turn23search1turn15search1turn21search9turn20search0turn20search5turn22search2turn22search7turn23search0turn18search2

The next commercial procurement should be **transactions and current listings**, not a giant “everything” package. Request proposals from PropTrack, Domain and QVAS suppliers around the exact fields required:

```text
Transactions
property identifier
sale date
sale price
transaction type
settlement/advice status

Listings
listing identifier
property identifier
status
listing date
withdrawn/sold/leased date
asking price/rent
property type
bed/bath/car
agency
agent
marketing text
permitted image references
```

PropTrack already exposes transaction, property, market and listing APIs; Domain exposes comparable agents/listings, property and valuation packages. citeturn23search3turn23search2

Then connect **first-party engagement**. REA's Campaign and Leads APIs are particularly important because they allow authorised campaign performance and enquiries to flow into the company's own data environment rather than remaining trapped in portal dashboards. citeturn17search1turn17search2

Only after those foundations are operating should the organisation spend heavily on predictive models.

The sequence I would use is:

| Delivery wave | Build |
|---|---|
| **Foundation** | Property master, address search, parcel maps, zoning, overlays, demographics, school/transport access |
| **Development intelligence** | DAs, development pipeline, competition, hazard screening and acquisition reports |
| **Market intelligence** | Licensed transactions/listings, rents, yields, comparable sales, supply and absorption |
| **Portfolio intelligence** | Leasing dashboard, valuations, rent-review opportunities, market-vs-portfolio performance |
| **Behavioural intelligence** | CRM + REA enquiries/campaign engagement |
| **Predictive layer** | Sale/listing propensity, forecasting, acquisition ranking and early-warning alerts |

The result is strategically different from simply subscribing to PropTrack.

PropTrack can remain an extremely valuable **input supplier**—its current platform offers more than a trillion property data points, commercial property information, transactions, listings, consumer signals and valuation products—but the company would own the data model, history, scoring and business logic. citeturn23search7turn23search11

That is the architecture I would choose for a Brisbane developer and leasing business: **do not attempt to recreate REA's network effects or PropTrack's national data factory; recreate the parts that are commoditised or publicly available, license the scarce raw data, and own every analytical layer above it.**

The competitive asset then becomes neither a data subscription nor a dashboard. It becomes a continuously improving **Brisbane property intelligence graph** capable of answering, for any address:

> **What is it? What can be built on it? What constraints apply? What is happening around it? What is it worth? What will it rent for? Who is competing nearby? How is demand changing? And is this site more attractive than every alternative currently under consideration?**

Queensland's address, cadastral, planning, transport, demographic, rental, crime, development and environmental datasets make most of that capability technically achievable in-house today; transactions, listings, ownership, valuations and portal behaviour are the relatively small—but commercially critical—licensed layer that completes the picture. citeturn19search1turn19search7turn23search1turn20search5turn22search2turn21search7turn22search7turn23search3