Field reference: News Signals

All 58 fields of the News Signals, grouped by category.

Every field may be null unless marked required. Today article_id is guaranteed.

Under each type is the column’s Parquet type, as DuckDB names it: VARCHAR is a string, BIGINT a 64-bit integer, DOUBLE a 64-bit float, and [] a list. Each category ends with its full Parquet schema, ready to copy.

What’s in it

Category	Fields
Base Article Info	13
Revenue	18
Funding	12
Deal	11
AI Scoring (with company context)	4

Base Article Info

Field	Type	Description
`article_id`	`integer` required `INTEGER`	Unique article id.
`article_url`	`string` `VARCHAR`	Original web address of the article or social media post.
`fetched_at`	`timestamp` `TIMESTAMP WITH TIME ZONE`	Timestamp at which the article or post was retrieved.
`title_original`	`string` `VARCHAR`	Title of the article as it was originally published in its native language.
`title_en`	`string` `VARCHAR`	English translation of the title.
`summary_en`	`string` `VARCHAR`	Summary in english.
`categories`	`string` `VARCHAR`	A set of thematic categories, delivered as a JSON array in a text column. One of 14 values, listed below the table.
`published_date`	`timestamp` `TIMESTAMP WITH TIME ZONE`	Date when the article or post was orginally published, as retrieved from the original source html.
`featured_image_url`	`string` `VARCHAR`	URL of the featured image of the article, if any.
`specter_id`	`string` `VARCHAR`	Specter ID of the linked company.
`company_name`	`string` `VARCHAR`	Name of the company the article is linked to.
`domain`	`string` `VARCHAR`	Domain of the company the article is linked to.
`news_traction`	`string` `VARCHAR`	Key traction highlights and figures mentioned in the source.

Parquet schema: Base Article Info

Base Article Info

(
  article_id          INTEGER,
  article_url         VARCHAR,
  fetched_at          TIMESTAMP WITH TIME ZONE,
  title_original      VARCHAR,
  title_en            VARCHAR,
  summary_en          VARCHAR,
  categories          VARCHAR,
  published_date      TIMESTAMP WITH TIME ZONE,
  featured_image_url  VARCHAR,
  specter_id          VARCHAR,
  company_name        VARCHAR,
  domain              VARCHAR,
  news_traction       VARCHAR
)

Accepted values for categories

Funding Announcement
Revenue Update
Growth
Strategic Partnerships & Collaborations
Market & Industry Trends
M&A
Decline
Product & Service Updates
Profitability Update
Regulatory & Legal Issues
Leadership & Governance Changes
Launch/Stealth
IPO and Exit Events
New VC Fund

Revenue

Field	Type	Description
`revenue_raw_revenue_amount`	`number` `DOUBLE`	The exact numerical revenue figure reported in the article.
`revenue_currency`	`string` `VARCHAR`	Three-letter currency code (e.g., EUR, USD) of the original revenue amount.
`revenue_metric`	`string` `VARCHAR`	The type of revenue stated. One of `REVENUE`, `ARR`, `GMV`.
`revenue_year`	`integer` `INTEGER`	Year to which the reported revenue figure refers.
`revenue_interval`	`string` `VARCHAR`	Period length (e.g., “quarterly”, “annually”) for the original revenue figure. One of `YEARLY`, `QUARTERLY`, `MONTHLY`, `HALF_YEARLY`, `WEEKLY`, `DAILY`.
`revenue_reference_date`	`timestamp` `TIMESTAMP WITH TIME ZONE`	The date as of which the stated revenue is valid. For example, EOY for yearly revenue, or a specific date for an ARR metric. Note this might differ from the publishing date. As an example, a tweet published in June might state “our ARR hit $5m at the beginning of the year.”.
`revenue_annualized_revenue_amount`	`number` `DOUBLE`	Annualized version of the raw revenue figure.
`revenue_annualized_revenue_amount_usd`	`number` `DOUBLE`	Annualized revenue figure, converted to USD.
`revenue_context`	`string` `VARCHAR`	Additional context about the reported revenue.
`revenue_tags`	`array<string>` `VARCHAR[]`	Processing tags to clarify the revenue number or how it has been processed. One of `COMPANY_REPORTED`, `ANNUALIZED`, `APPROXIMATE`, `PROJECTION`, `ESTIMATED`, `RANGE`, `RECURRING_REVENUE`.
`revenue_processing_notes`	`string` `VARCHAR`	Model notes explaining how revenue data was processed or assumptions made.
`revenue_source`	`string` `VARCHAR`	Source stated as text (mostly the name of the publication)
`revenue_is_profitable`	`boolean` `BOOLEAN`	True if the article states the company was profitable during the reported period.
`revenue_annual_profit_usd`	`number` `DOUBLE`	Annualized profit figure (if available), converted to USD.
`revenue_annual_profit_year`	`integer` `INTEGER`	Year to which the reported profit figure refers.
`revenue_profitability_context`	`string` `VARCHAR`	Additional context about the reported profit.
`revenue_publish_date`	`timestamp` `TIMESTAMP WITH TIME ZONE`	When the revenue was published (should mostly align with published_date)
`revenue_profitability_reference_date`	`timestamp` `TIMESTAMP WITH TIME ZONE`	The date as of which the stated profitability metric is valid.

Parquet schema: Revenue

Revenue

(
  revenue_raw_revenue_amount             DOUBLE,
  revenue_currency                       VARCHAR,
  revenue_metric                         VARCHAR,
  revenue_year                           INTEGER,
  revenue_interval                       VARCHAR,
  revenue_reference_date                 TIMESTAMP WITH TIME ZONE,
  revenue_annualized_revenue_amount      DOUBLE,
  revenue_annualized_revenue_amount_usd  DOUBLE,
  revenue_context                        VARCHAR,
  revenue_tags                           VARCHAR[],
  revenue_processing_notes               VARCHAR,
  revenue_source                         VARCHAR,
  revenue_is_profitable                  BOOLEAN,
  revenue_annual_profit_usd              DOUBLE,
  revenue_annual_profit_year             INTEGER,
  revenue_profitability_context          VARCHAR,
  revenue_publish_date                   TIMESTAMP WITH TIME ZONE,
  revenue_profitability_reference_date   TIMESTAMP WITH TIME ZONE
)

Funding

Field	Type	Description
`funding_amount`	`number` `DOUBLE`	Amount of funding raised by the company, as reported.
`funding_amount_usd`	`number` `DOUBLE`	Funding amount converted to USD.
`funding_currency`	`string` `VARCHAR`	Three-letter currency code for the original funding amount.
`funding_round`	`string` `VARCHAR`	Specific funding round mentioned in the article. Note this is free text, not an enum.
`funding_date`	`date` `DATE`	Date when the funding announcement was made or round closed.
`funding_investors`	`array<string>` `VARCHAR[]`	List of investor names mentioned as participants in the funding round.
`funding_matched_investor_ids`	`array<string>` `VARCHAR[]`	Specter ids uniquely identifying investors linked to this funding round.
`funding_notes`	`string` `VARCHAR`	Any additional context or details about the funding structure or its intended use.
`funding_valuation`	`number` `DOUBLE`	Post-money valuation of the company stated in relation to the funding round.
`funding_valuation_usd`	`number` `DOUBLE`	Company’s post-money valuation converted to USD.
`funding_valuation_currency`	`string` `VARCHAR`	Three-letter currency code for the original funding valuation.
`funding_is_breaking_news`	`boolean` `BOOLEAN`	True if the article reports the funding event as new information, false if mentioning past.

Parquet schema: Funding

Funding

(
  funding_amount                DOUBLE,
  funding_amount_usd            DOUBLE,
  funding_currency              VARCHAR,
  funding_round                 VARCHAR,
  funding_date                  DATE,
  funding_investors             VARCHAR[],
  funding_matched_investor_ids  VARCHAR[],
  funding_notes                 VARCHAR,
  funding_valuation             DOUBLE,
  funding_valuation_usd         DOUBLE,
  funding_valuation_currency    VARCHAR,
  funding_is_breaking_news      BOOLEAN
)

Deal

Field	Type	Description
`deal_type`	`string` `VARCHAR`	Nature of the transaction (e.g., “strategic_partnership”, “acquisition_of_company”) One of `strategic_partnership`, `acquisition_of_company`, `acquisition_by_company`, `asset_sale`, `merger`.
`deal_acquisition_price`	`number` `DOUBLE`	Stated acquisition price of the deal as reported.
`deal_acquisition_price_usd`	`number` `DOUBLE`	Acquisition price of the deal converted to USD.
`deal_acquisition_currency`	`string` `VARCHAR`	Three-letter currency code for the original deal acquisition amount.
`deal_date`	`date` `DATE`	Date when the deal was officially signed or publicly announced.
`deal_context`	`string` `VARCHAR`	Narrative description or additional context about the deal details.
`deal_counterparty`	`string` `VARCHAR`	Name of the other party involved in the deal.
`deal_is_breaking_news`	`boolean` `BOOLEAN`	True if the article reports the deal as new information, false if mentioning past.
`deal_matched_specter_ids`	`array<string>` `VARCHAR[]`	Specter id(s) uniquely identifying the counterparty involved in the deal.
`deal_status`	`string` `VARCHAR`	Current status of the deal (e.g., “announced”, “completed”, “pending”) One of `announced`, `completed`, `pending regulatory approval`, `pending`, `rejected`, `approved`.
`deal_structure`	`string` `VARCHAR`	Additional details about the structure of the deal (e.g., “majority stake”, “asset purchase”)

Parquet schema: Deal

Deal

(
  deal_type                   VARCHAR,
  deal_acquisition_price      DOUBLE,
  deal_acquisition_price_usd  DOUBLE,
  deal_acquisition_currency   VARCHAR,
  deal_date                   DATE,
  deal_context                VARCHAR,
  deal_counterparty           VARCHAR,
  deal_is_breaking_news       BOOLEAN,
  deal_matched_specter_ids    VARCHAR[],
  deal_status                 VARCHAR,
  deal_structure              VARCHAR
)

AI Scoring (with company context)

Field	Type	Description
`score`	`integer` `INTEGER`	Score indicating the significance of the news update for the company.
`reasoning`	`string` `VARCHAR`	Explanation of the assigned significance score.
`meaningful_updates`	`string` `VARCHAR`	Types of meaningful updates (w.r.t. To the context of the company) present in the article, delivered as a JSON array in a text column. One of `REVENUE_UPDATE`, `DEAL_UPDATE`, `TRACTION_UPDATE`, `FUNDING_UPDATE`, `PROFIT_UPDATE`.
`whats_new`	`string` `VARCHAR`	Concise summary of significant new datapoints compared to previous company context.

Parquet schema: AI Scoring (with company context)

AI Scoring (with company context)

(
  score               INTEGER,
  reasoning           VARCHAR,
  meaningful_updates  VARCHAR,
  whats_new           VARCHAR
)

Get started

Core data

Signals

Field reference: News Signals

What’s in it

Base Article Info

Revenue

Funding

Deal

AI Scoring (with company context)

​What’s in it

​Base Article Info

​Revenue

​Funding

​Deal

​AI Scoring (with company context)

What’s in it

Base Article Info

Revenue

Funding

Deal

AI Scoring (with company context)