Facets
The label vocabulary for domains, impact, region, relevance, sentiment, and tone.
Domains
Each article is tagged with one or more domains:
geopolitics_conflict, tech_semis, equities_earnings, trade_supply, macro_growth,
energy, banking_credit, monetary_rates, commodities_metals, crypto,
housing_realestate, other.
Filter with ?domain=crypto to match articles tagged with that domain.
Impact
low / medium / high: the model's estimate of market impact. Filter with ?impact_min=high
(returns high only; ?impact_min=medium returns medium and high).
Region
Where the story matters (e.g. US, Global). Filter with ?region=US.
Relevance
relevance (0-1) is the gate's confidence that the article is market-relevant. Only articles that
pass the gate are stored, so every record is relevant, and the score tells you how strongly.
Sentiment
sentiment is a FinBERT score (higher = more positive); sentimentLabel is the class
(positive / negative / neutral). Filter with ?sentiment=positive.
Tone
tone is the raw GDELT tone score for the source article, a signed number where higher is more
positive. It comes straight from GDELT and is separate from the FinBERT sentiment above. It is
returned on every record but is not a filter.