Data integrity is part of the product.
Signavaris is designed to preserve the difference between source evidence, transformed measures, derived intelligence, projections and research signals. A clean-looking chart is not enough: the evidence behind it must remain traceable.
Rights before use
Sources, datasets and indicators must pass the applicable Signavaris rights and governance gate before ingestion, analysis, derivation or customer exposure. Material that does not meet the required rights status is not used.
Provenance retained
Where supported, observations retain provider, dataset, indicator, geography, period, unit and frequency. Signavaris-derived measures are identified separately from source observations.
No invented observations
Signavaris does not fabricate missing raw data. Source gaps remain gaps unless a clearly identified analytical method legitimately derives a separate measure.
Validation before promotion
Correlations and candidate signals are not automatically treated as forecasts. Research can be tested for multiple comparisons, time ordering, out-of-sample behaviour, cross-country transfer, placebo performance, regime stability and independent replication where appropriate.
Failure is evidence
If independent evidence weakens a result, its status should be downgraded rather than hidden. Models and interpretations may be corrected, reclassified or withdrawn as evidence changes.
Public sources, Signavaris intelligence
Some underlying evidence originates from third-party public institutions and remains attributed to them. Signavaris value lies in governance, harmonisation, comparison, validation, evidence routing and decision-focused analysis built across those sources.