Methodology · the radar engine

How we score it

Construction and land are ordered from public evidence. Commercial offers are a separate lane: we show their published data and separate what has a linked public cross-check from what it still does not prove. We do not tell you whether to buy: we reduce information asymmetry.

Methodology reviewed on September 4, 2026 · v3

Reading modelThe signal is separated before it is ranked

01
SourceAuthority · document · date
02
CoverageMeasured · declared · missing
03
Risk + chronologyTwo axes, never a promise

OutputTraceable record

The index helps rank later; it never fills gaps or offsets a critical flag.

Citable definitions

These terms keep the same meaning across signal pages, offers, zones, and reports.

MADES signal: A public environmental record observed by OJOPY. It is not a municipal permit, construction start, availability, or sale.
Commercial match: An attributable, dated listing linked to the same identity. Without a verifiable match, the commercial chronology remains undetermined.
Pairing: Visible price and area that belong to the same published unit. OJOPY does not combine operands from different units.
Benchmark: A cohort of distinct commercial projects with the same asset type and similar area in the narrowest viable radius. A weak or dispersed cohort is withheld.
Confidence: A label for coverage and quality of the measured data. It is not a probability that the project will succeed.
Freshness: A state derived from absolute source and check dates. A historical reference is not presented as current availability.
Actionability: A next review step produced by evidence and risk rules. It is never a buy or sell recommendation.
Public activity: Filings or publications observed in identified sources. It does not establish deliveries, sales, or completed investment.

1 · What we measure

Each signal is worked through in layers. This is not an appraisal or a legal opinion: it is a map of what was observed in identified sources, its limits, and the data that remains missing.

The projectWhat it is, what is promised, what stage it is at, and what public signal surfaced it (environmental filing in MADES/SIAM, registration, announcement).

The locationArea, surroundings, access, and observed listing and comparable context. It does not predict demand or how easily this asset could be resold or rented.

The developer / entityWho is behind it, what public track record they have, and whether the entity is on record and identifiable in public registries.

Public signalsFilings, permits, announcements, and digital trail that confirm (or contradict) what is promised.

Price / m² and comparablesHow the price compares against similar properties in the same area.

Public and missing documentationWhat appears in public sources, what should exist, and what is conspicuously absent.

Yield promisesWhich inputs the promise discloses and how they compare with available rental evidence; OJOPY does not validate a future return.

Red flagsObserved inconsistencies, gaps, and conditions that require legal, construction, delivery, price, or liquidity review.

QuestionsThe specific questions worth taking to the seller, the lawyer, or the notary.

2 · How we score it

We cross-reference public sources with comparable data from the area. The engine organizes scattered information, measures coverage and risk over public sources, and marks what is missing and what is left without a public source to back it.

What matters is not the reading alone, but its provenance. The Evidence Brief identifies source and date when available, distinguishes derived calculations, and keeps fields without attributable support explicit.

We cross-referencepublic sources + local price and rental comparables.

The engine organizesgroups signals, weighs the risk, and marks what is missing — the verdict is algorithmic, never a buy recommendation.

Available provenancewe show source and date when recorded; if either is missing or a reading is derived, we say so instead of filling the gap.

2b · Two timing states, not one

OJOPY does not use “Timing” as a label that mixes age with commercial presence. The MADES date and chronology against a listing answer different questions and are published separately.

Signal recencyMeasures how long it has been since the usable MADES source date: very recent, recent, mature, old, or date unavailable.

Commercial matchOnly compares MADES with an attributable, dated listing: undetermined, confirmed lead, already public, or coincident.

Fail-closed ruleWithout an attributed, dated listing, we claim neither a lead nor that the project was already public. Area comparables do not replace that match.

3 · How to read the index

The 0-100 index consistently ranks signals inside OJOPY. It is ordinal: it is not a probability of success, an appraisal, an expected return, or a calibrated scientific measurement.

A small difference — for example, 62 versus 58 — should not decide anything on its own. Read the grade and band first; then review which data is covered, what is missing, and which sources support each finding.

Public viewshows bands by dimension to avoid context-free apparent precision.

Evidence Briefshows the exact index alongside its drivers, sources, and missing data.

Model changesa recalibration can move the index without any project or market change; the history identifies it separately and restarts the comparison.

3b · How to read commercial offers

A commercial offer is not a construction record: it starts from an observed commercial listing. We group it only when name, developer and address/area match exactly, show the data it publishes and do not reuse the radar's index or S–F letters.

Commercial sheetWe state how many of six basic fields are published: name, developer, address, type, price and surface. The count describes data availability, not a complete project file or a rating.

Public cross-checkWe mark a cross-check as linked only when a verifiable match exists. We state whether the MADES file is administrative, construction-related or still unclassified: an audit or ownership change does not corroborate construction or lift the sheet. “Not linked yet” is work pending at OJOPY, not a claim that no record exists.

Declared stageWe show pre-sale, off-plan, under construction, immediate delivery or completed only when the commercial source states those words, and preserve the source phrase. We do not infer a stage from a numeric identifier.

What it does not claimWe do not infer permits, construction progress, delivery, solvency, contract, returns or whether you should buy. Those checks remain pending even when the card has six published commercial fields.

DatesWe distinguish the date added to OJOPY from OJOPY's last check. We do not show a portal publication or update date unless we have it as source data.

4 · Weights, missing data, and penalties

Weights are relative within each asset type. For construction, the approximate active mix is: location 21%, index timing context 19%, rental context 16%, developer 15%, market traction 10%, public documentary trail 10%, liquidity 5%, and saturation 4%.

For land: location 26%, land pricing 20%, documentary trail 16%, liquidity 12%, saturation 10%, index timing context 8%, traction 5%, and subdivider 3%. That numeric contribution does not replace the two public states above. Rental and construction timing do not score raw land.

Unknown datais null, not zero. It is not treated as bad and is never filled with an assumption.

Effect of missing datathe calculation renormalizes over measured dimensions, but coverage and confidence fall; below 50% coverage the grade is capped, and below 30% it cannot exceed D.

Risk pressurethe worst observed risk leads, with a fraction of other relevant risks added. At 75/100 the signal moves to critical review.

Versionsthe current model is v3. If weights or rules change, history marks the recalibration and resets the comparison so it is not attributed to the project.

Historical validationwe publish retrospective validations against dated sources, including counterexamples and matching failures. We do not yet present the index as a calibrated predictive model.

5 · Risk scale

Each block of the Evidence Brief carries a risk level. What could not be verified is not assumed to be fine: it is flagged as insufficient information.

LowThe rules detect no relevant conflicts in covered fields; missing data remains unknown.

MediumThe rules flag points worth clarifying; the band neither clears nor blocks a decision.

HighThe rules flag material issues that require evidence and professional review.

CriticalThe rules detect a critical contradiction or risk pressure; prioritise its review before deciding.

Insufficient informationCoverage is not enough for a reading; request documentation and keep the gaps marked as unknown.

6 · Preliminary reading

We close with one of five reading states. They describe the result of rules applied to covered data and its gaps; they never mean "buy" or "don't buy".

MonitorCoverage or dated evidence is still limited; the next step is to monitor changes.

Request evidenceData or attributable sources are missing for the flagged gaps.

CompareComparables are available to contrast the signal, with their limits and dispersion visible.

Investigate furtherDetail is available for reviewing sources, gaps, and risks; it does not amount to approving the project.

Critical reviewThe rules flag issues or risk pressure that should be reviewed before the rest of the reading.

The reading organises evidence and gaps; it neither certifies the project nor recommends buying.

7 · The honest limit

It does not replace a lawyer or a notary: it makes the buyer better prepared, it does not replace legal review.
It does not guarantee that the project is safe or that the investment will work out.
We are not a broker and we do not sell properties.
It aims to reduce information asymmetry by making risks, limits, and questions visible before a decision.
An environmental date may precede or follow marketing; a timing advantage is only stated against a specific dated source.

Open the radar See offersChoose public evidence for construction and land; use Offers to research commercial inventory with its limits in view.

Want to see the output? Audited real case · pricing · why we are independent.

Research graph

Six related routes, selected by function rather than volume.

01Explore construction and land signalsPublic records, dates, coverage, and gaps for each signal.02Cross-check observed commercial offersPublished price and area with visible sources and limitations.03Compare zones with dated evidenceDistribution, freshness, and contexts that may remain unmeasured.04Review the due-diligence guideWhich document to request for each claim before committing.05Read the audited case and its controlsVerifiable chronology, counterexamples, and report boundaries.06Prepare a purchase from abroadTitle, cadastre, permits, and the border-zone exception with sources.