How does Casa Daily score a property?

Each property gets a 0-100 Casa Score composed of seven weighted components: zone permissibility (what the planning scheme actually allows), lot characteristics (area, frontage, depth, shape, slope), overlay impact (flood, bushfire, character, heritage, vegetation, koala), DA precedent (what the council has approved within 500m on similar lots), market conditions (comparable sales and end-product values), use diversity (how many viable typologies the lot can carry), and assessment complexity (Code vs Impact, IR risk, specialist trigger count).

Where does Casa Daily get its data?

Six primary sources: (1) council planning scheme APIs and PDF schedules for zone, sub-precinct, and overlay maps; (2) council DA registers — over 800,000 historical applications across SEQ, weekly refresh; (3) Department of Resources cadastral data for lot dimensions and ownership; (4) REA + Domain for live listing data, refreshed multiple times per day; (5) settled-sales registers for comparable end-product values; (6) infrastructure charges schedules published by each council.

How accurate is the Casa Score?

The Casa Score is a deterministic ranking signal — it doesn't predict approval outcome with confidence intervals because no single number can. It uses ordinal ranking (a 78 is a better-quality opportunity than a 71) rather than absolute prediction. The intelligence layer beneath the score (zoning, overlays, DA history, comparables) is pulled from authoritative sources and is the same data a town planner would research.

Does Casa Daily use AI to score?

Casa Daily uses deterministic rule-based scoring for the seven components — zoning rules, overlay rules, lot geometry rules, comparable approval matching are all hand-coded against the published planning schemes. AI (Claude) is used in two specific places: (a) the conversational analyst beside every listing, with the property data and council corpus loaded; (b) the document-checker that compares uploaded planning reports against council requirements. Scoring itself is rules-driven, not LLM-generated.

Why deterministic scoring rather than ML?

Three reasons. One: planning rules are explicit, published, and rule-based — replicating them deterministically gives perfect interpretability. Two: training data for ML scoring would require thousands of paired (listing, eventual development outcome) records — and outcomes are years away from each listing. Three: when a Casa Score is wrong, we can trace exactly which component drove it; ML scores fail opaquely. We treat AI as a layer ABOVE the deterministic data, not a substitute for it.

Methodology

How Casa Daily scores SEQ property development potential.

Casa Daily produces a 0-100 Casa Score for every fresh REA listing across South-East Queensland. The score is a weighted composite of seven deterministic components — every input traceable to a published source, no black box. This page documents the method in full.

The seven scoring components

Each component contributes a 0-100 sub-score. The Casa Score is a weighted average, with weights tuned per opportunity type (subdivision-specific weights differ from townhouse-specific weights, etc.).

Component	What it measures	Source
Zone permissibility	What the planning scheme actually allows in this zone and sub-precinct (single dwelling, dual occupancy, multiple dwelling, secondary dwelling, etc.).	Council planning scheme + Categories of Development tables.
Lot characteristics	Area, frontage, depth, shape, slope, access. Whether the geometry can carry the proposed dwelling pattern.	Department of Resources cadastral data.
Overlay impact	Flood, bushfire, traditional building character, heritage, vegetation protection, koala habitat, landslide, coastal, environmental.	Council overlay maps (live).
DA precedent	What the council has actually approved within 500m on similar lots, in the last 36 months.	Council DA registers (800,000+ records).
Market conditions	Comparable settled sales, end-product price points, current days-on-market for the suburb's typology.	Settled-sales registers + REA/Domain market data.
Use diversity	How many viable development typologies the lot can carry. A lot that supports subdivision, dual occupancy AND a granny flat scores higher than one that supports only one.	Cross-product of zone permissibility × lot characteristics.
Assessment complexity	Code vs Impact assessment level, expected Information Request risk, specialist trigger count (acoustic, ecological, traffic).	Council Categories of Development + recent IR patterns.

What a Casa Score number means

Score	Meaning
90+	Rare. Worth aggressive action — typically the binding constraint is small and the upside is multi-typology.
80–89	Strong. Worth an inspection. Multiple opportunity types likely viable; constraints minor and known.
70–79	Genuinely interesting. Worth opening the dashboard for full Snapshot intelligence.
60–69	Marginal. One or two binding constraints; works for some strategies, not others.
50–59	Below threshold. Significant constraint (overlay, zoning, geometry) reduces upside meaningfully.
<50	Skip. The data shows the proposed development pathway is materially blocked.

Note: the Casa Score is a ranking signal, not an approval prediction. A 78 represents a higher-quality opportunity than a 71, but neither is a guarantee of council approval. Confirm with a registered town planner before committing capital.

Data sources, verbatim

We don't hide the sources. Every component above traces to a publicly-citable origin:

Brisbane City Council: City Plan 2014, Strategic Framework, PD Online DA register, infrastructure charges schedule.
City of Gold Coast: City Plan 2016, CityPlan Online maps, DA Tracker.
Sunshine Coast Regional Council: Sunshine Coast Planning Scheme 2014.
City of Moreton Bay: Moreton Bay Regional Council Planning Scheme.
Logan City Council: Logan Planning Scheme 2015, Logan PD Hub.
Redland City Council: Redland City Plan.
Ipswich City Council: Ipswich Plan.
Cadastral data: Queensland Department of Resources DCDB.
Listings: realestate.com.au and domain.com.au listing pages.
Settled sales: Australian Bureau of Statistics + state property registers.

Methodology FAQ

How does Casa Daily score a property?: Each property gets a 0-100 Casa Score composed of seven weighted components: zone permissibility (what the planning scheme actually allows), lot characteristics (area, frontage, depth, shape, slope), overlay impact (flood, bushfire, character, heritage, vegetation, koala), DA precedent (what the council has approved within 500m on similar lots), market conditions (comparable sales and end-product values), use diversity (how many viable typologies the lot can carry), and assessment complexity (Code vs Impact, IR risk, specialist trigger count).
Where does Casa Daily get its data?: Six primary sources: (1) council planning scheme APIs and PDF schedules for zone, sub-precinct, and overlay maps; (2) council DA registers — over 800,000 historical applications across SEQ, weekly refresh; (3) Department of Resources cadastral data for lot dimensions and ownership; (4) REA + Domain for live listing data, refreshed multiple times per day; (5) settled-sales registers for comparable end-product values; (6) infrastructure charges schedules published by each council.
How accurate is the Casa Score?: The Casa Score is a deterministic ranking signal — it doesn't predict approval outcome with confidence intervals because no single number can. It uses ordinal ranking (a 78 is a better-quality opportunity than a 71) rather than absolute prediction. The intelligence layer beneath the score (zoning, overlays, DA history, comparables) is pulled from authoritative sources and is the same data a town planner would research.
Does Casa Daily use AI to score?: Casa Daily uses deterministic rule-based scoring for the seven components — zoning rules, overlay rules, lot geometry rules, comparable approval matching are all hand-coded against the published planning schemes. AI (Claude) is used in two specific places: (a) the conversational analyst beside every listing, with the property data and council corpus loaded; (b) the document-checker that compares uploaded planning reports against council requirements. Scoring itself is rules-driven, not LLM-generated.
Why deterministic scoring rather than ML?: Three reasons. One: planning rules are explicit, published, and rule-based — replicating them deterministically gives perfect interpretability. Two: training data for ML scoring would require thousands of paired (listing, eventual development outcome) records — and outcomes are years away from each listing. Three: when a Casa Score is wrong, we can trace exactly which component drove it; ML scores fail opaquely. We treat AI as a layer ABOVE the deterministic data, not a substitute for it.

See the Casa Score in action. First brief tomorrow morning.

Create your account

$500 / mo · or $4,800 / yr (save 20%) · cancel any time