Independent thinking. Connected markets.Our approach to intelligence

DATA QUALITY

Measure freshness at the source, not at the screen

A practical way to separate collection health, observation age and data completeness.

Three clocks, three meanings

An observation time says when the underlying price or condition was observed. A collection time says when a system retrieved it. A rendering time says when the interface was generated. They may differ substantially, so each belongs in its own field. Updating the rendering time does not make the underlying observation fresh.

Define validity by the feed contract

A market-rate feed, an institutional bulletin and a provider’s quoted transfer price can have different validity rules. Use the source contract and the analytical use case to set a validity boundary. Avoid a universal “live” badge across datasets that are updated on different schedules.

Make failure visible

A failed refresh should not replace a valid timestamp with the current time. If stale data are retained for historical context, identify them clearly. Where an expired price would mislead a customer, withhold the actionable amount and show the reason.

A compact evidence record

Useful observation metadata include the source, observation time, collection time, validity boundary, currency pair, transaction conditions and a publication-rights decision. These fields let analysts distinguish a collection problem from a genuinely changing market.

LET’S START WITH YOUR QUESTION

A clearer view.
A better next move.

Tell us what you need to understand. We’ll start there.

Talk to usGet to know the platform ↗