Entity resolution
How Datahyena collapses scattered mentions into one canonical company, investor, and event.
The same company shows up across dozens of outlets, spelled a dozen different ways, and the same funding round gets reported by everyone at once. Datahyena collapses all of that into one clean record. This is the work that makes the data joinable.
One canonical company
Every signal resolves to a single canonical company with a stable companyId, regardless of how the underlying reports spelled the name or which domain they linked. Investors and people are resolved the same way. So:
- Filter any endpoint by
companyIdand you get that company across funding rounds, acquisitions, and executive moves. - The
companyIdis stable, so it is safe to store as a join key against your own CRM or warehouse. - The same investor is the same
investors[].ideverywhere it appears, so you can track an investor's activity across deals.
companyId is a durable identifier. Persist it once and reuse it to join every future signal
about that company, rather than matching on names or domains yourself.
One event, not twelve rows
When a single real-world event is reported by many outlets, Datahyena returns one event, not one row per outlet. That event carries the full list of sources that reported it in its sources[] array, each with a URL. So corroboration is visible: an event confirmed by many outlets reads differently from one a single source mentioned, and you never double-count the same round.
What you can rely on
- Stable ids for companies, investors, and events, safe to persist and join on.
- No duplicates for the same real-world event.
- Visible corroboration through the
sources[]on every record.
See Coverage for how many entities are resolved today, and the data lifecycle for when a resolved event becomes available.