EVIDENCE · SOURCE, STATE AND HASH · EARLY ACCESS
Evidence with source, state and hash.
Test the decision with declared synthetic data, read elections from the official files of Brazil's electoral court (TSE) and calibrate with Brazilian public series. Every number states where it came from, what state it is in and which file supports it.
Studio and Scenarios run in your browser, no account and no personal data. Every generated record is labelled as synthetic.
| oportunidade_id | origem | valor_brl | dias_no_funil | converteu |
|---|---|---|---|---|
| OPP-00001 | parceiros | 8,023.39 | 19 | true |
| OPP-00002 | outbound | 12,988.16 | 15 | false |
| OPP-00003 | inbound | 12,010.51 | 12 | false |
| OPP-00004 | inbound | 36,076.91 | 16 | false |
| OPP-00005 | inbound | 35,934.95 | 17 | false |
- Validity
- 100%
- Fit
- 97.3%
- Rows in groups smaller than 5
- 0%
dias_no_funil · lognormal(3.3, 0.6)
NUMBERS WITH A SOURCE
What is published, counted when this page was published.
No number in this section was typed: all of them come from the manifests and the Studio engine.
- Official files with SHA-256
- 78
- 67 from the TSE · 11 series and tables from BCB, IBGE, Ipea and the Treasury
- Observed Source: TSE manifests and public sources catalog · latest collection on
- Published election cases
- 2
- Presidential 2022 and 2026, with comparison between cycles
- Observed Derived Source: elections index, with each case manifest ·
- Templates privacy-tested on every publication
- 4
- 2 on the k-anonymity target (k ≥ 5) · 1 below target, declared in the report · 1 with no quasi-identifier marked
- Synthetic Measured when this page was published, on synthetic data · method
WHAT YOU CAN DO TODAY
Choose where to start.
Each door shows the state of the evidence you will find.
- 01 · STUDIOGenerate a synthetic datasetBusiness template or CREATE TABLE, seed, validity, fit and privacy report, export with a manifest.SyntheticOpen
- 02 · SCENARIOSTest a decisionAssumptions by segment become thousands of scenarios, with intervals, sensitivity and what the decision supports.SyntheticOpen
- 03 · ELECTIONSRead the 2022 and 2026 casesResults, turnout, margin and swing by state from the TSE files, with the comparison between cycles.Observed DerivedOpen
- 04 · PUBLIC SOURCESCalibrate with official dataSeries from BCB, IBGE, Ipea and the Treasury downloaded unchanged, with SHA-256, as a reference in the Studio.ObservedOpen
- 05 · POLITICAL POSITIONMeasure your own positionOur own instrument, with a published method. Computed in your browser; nothing goes to the server without your action.DerivedOpen
- 06 · CLAIMSCheck a claimEvery claim with the state the collected evidence supports and the reason, including when it is unknown.UnknownOpen
SIX STATES, ALWAYS VISIBLE
Every number declares where it comes from.
The chain is always the same: file, observation, calculation, claim. A number never changes state by insistence; it only changes when a new source is collected and checked.
- Observed
- Read from a primary official file, with URL, download date and SHA-256.
- Imported
- Comes from a secondary source that republishes the data. A lead, not a primary fact.
- Derived
- A calculation on observed numbers, with the formula written next to it.
- Inferred
- Estimated by a method from other data. Not an observation.
- Synthetic
- Generated from declared assumptions, with a seed. It represents no real person or company.
- Unknown
- No collected source supports the item. The reason is written; unknown is not false.
GOVERNANCE IN EVERY FILE
Every export carries a manifest.
Seed, generator version, specification, metrics and limits travel with the CSV. Whoever receives the file knows where it came from and what it does not prove.
With public data, the same rule: the 2026 Brazilian presidential election case carries the SHA-256 of every TSE file, states what is observed, derived or unknown and ends with a ready-to-use citation.
{
"rotulo": "DADOS SINTÉTICOS",
"gerador": "Reveno Evidence · Estúdio",
"semente": 11,
"linhas": 1200,
"sha256_csv": "…",
"metricas": { "validade": 1, "aderencia": 0.98 },
"limites": [
"Não representa nenhuma pessoa ou empresa real.",
"Fidelidade a dados reais não é medida."
]
}
WHAT SYNTHETIC DATA IS
Data shaped like the business, with no real person in it.
A generator creates new records that follow declared distributions and relationships. It lets you test ideas, workflows and agents before touching the real base. What matters is knowing where it comes from.
- Simulated from assumptions
- You declare columns, distributions and correlations. This is what the Studio does today, in the browser.
- Available
- Synthetic from real data
- A model learns from your base and generates new records, with fidelity and risk reports.
- Early access
WHAT THIS DOES NOT PROVE
Knowing the limit is part of the evidence.
- Data generated from assumptions does not estimate the market or replace research with real people.
- Fit measures fidelity to the specification, not to real data.
- Privacy checks are not an anonymization opinion.
- An election case describes the published files: it makes no forecast, no voting recommendation and no judgement of candidates.
- A synthetic scenario is not a forecast: it shows what the assumptions support. Read the method
EARLY ACCESS
Want to generate from your own data?
Synthetic from real data, saved studies and method review join the platform through early access.