Event Lifecycle
The event lifecycle is the central operating process of NataPulse.
1. Collection
Section titled “1. Collection”An enabled provider returns a source item. NataPulse records the provider, source identifier, retrieval time, and available provenance.
2. Validation
Section titled “2. Validation”The pipeline checks that the item is structurally usable. Invalid, malformed, unsupported, stale, or prohibited inputs can be rejected or marked for review.
3. Normalization
Section titled “3. Normalization”Provider-specific fields are mapped into a common event structure. Text is cleaned for display, timestamps are standardized, and structured domain fields are preserved.
4. Entity resolution
Section titled “4. Entity resolution”The system attempts to identify the company, ticker, crypto asset, wallet, protocol, sector, person, or topic involved. Ambiguous names are not blindly converted into tickers.
5. Deduplication
Section titled “5. Deduplication”Provider identifiers, content fingerprints, canonical links, timing, entities, and other signals prevent repeated ingestion from becoming repeated evidence.
6. Classification and scoring
Section titled “6. Classification and scoring”The event receives source, relevance, importance, confidence, and other applicable classifications. Scores are used for prioritization and disclosure, not as guarantees.
7. Publication gate
Section titled “7. Publication gate”The event is checked against product publication requirements. Low-quality or unsupported material may remain internal or be excluded. Social claims from unverified sources receive stricter treatment when uncorroborated.
8. Clustering
Section titled “8. Clustering”Related published events can join a stable cluster based on canonical entity, timing, and relationship signals. A persistent cluster generally represents multiple related observations; a real-time interface may temporarily show a new item before corroboration arrives.
9. Narrative derivation
Section titled “9. Narrative derivation”Clusters and events that persist, broaden across sources, or accelerate can contribute to an emerging narrative.
10. Downstream use
Section titled “10. Downstream use”Published events can trigger or support reports, Deep Research, watchlist rules, alerts, cited chat answers, market-impact analysis, and specialist-desk views.
11. Outcome and memory
Section titled “11. Outcome and memory”Later evidence can confirm, partially confirm, invalidate, or leave a research conclusion inconclusive. Safe lessons may be stored and retrieved in future work on the same scope.
An event can therefore move from a single observation to part of a larger, historically evaluated chain of evidence.