Skip to main content
This section introduces the main concepts behind the ViceWire data model. The goal is not to document every field in one place. The goal is to explain how ViceWire thinks about company-specific events, repeated coverage, sentiment, relevance, geography, and claim certainty so the rest of the payload is easier to interpret.

Event Model Overview

How one article becomes one or more structured event objects for a specific security.

Directional Layers

How event meaning and FinBERT tone are kept as separate but complementary layers.

Business Impact Mapping

How each event is mapped to the part of the company most directly implicated.

Novelty Tracking

How ViceWire separates genuinely new developments from repeated, syndicated, and related follow-up coverage.

Entity Relevance Signals

How article-level signals distinguish true company coverage from passing or peripheral mentions.

Geography Data

How place is captured only when it is materially relevant to the event described in the article.

Facts vs Rumors

How ViceWire separates officially confirmed claims from sourced reporting and market speculation.
The current public examples in these docs focus on Apple, but the concepts in this section describe the underlying company-specific event model more broadly.