Datasetpublic

Scottswi/fenrix-financial-news-lake

FENRIX Financial News Lake Topic-organized financial news data for portfolio-management decision research. The verified sanitized backing lake currently contains 37,571,131 rows across 1,461 Parquet shards and 8 sources. The product layer is organized by article meaning, not raw source lineage: article_group -> article_type -> year/month -> rows Source, license, URL, and provenance fields remain row-level audit metadata. Recommended Loading from datasets import… See the full description on the dataset page: https://huggingface.co/datasets/Scottswi/fenrix-financial-news-lake.

sourceHugging Faceotherupdated 3mo agoView on Hugging Face

1likes654downloads

Dataset Card

FENRIX Financial News Lake

Topic-organized financial news data for portfolio-management decision research.

The verified sanitized backing lake currently contains 37,571,131 rows across 1,461 Parquet shards and 8 sources. The product layer is organized by article meaning, not raw source lineage:

text

article_group -> article_type -> year/month -> rows

Source, license, URL, and provenance fields remain row-level audit metadata.

Recommended Loading

python

from datasets import load_dataset

sample = load_dataset("Scottswi/fenrix-financial-news-lake", "sample_pm_decision_news")
macro = load_dataset("Scottswi/fenrix-financial-news-lake", "economic_macro", streaming=True)
single_stock = load_dataset("Scottswi/fenrix-financial-news-lake", "single_stock", streaming=True)
sector = load_dataset("Scottswi/fenrix-financial-news-lake", "sector_industry", streaming=True)

Filter after load:

python

rows_2020 = (row for row in sample["train"] if row["year"] == "2020")
aapl = (row for row in sample["train"] if "AAPL" in row.get("tickers", "[]"))

Gold Schema

Important product fields:

—article_group
—article_type
—topic
—subtopic
—decision_trigger
—decision_context
—market_scope
—published_at
—title
—description
—body
—tickers
—companies
—sectors
—macro_tags
—materiality_score
—novelty_score
—reverse_id_risk_score

Important audit fields:

—source_id
—publisher
—canonical_url
—license_class
—redistribution_status
—provenance_uri
—raw_payload_uri
—dedupe_key

Current Status

—Sanitized silver verification: reports/news_lake_sanitized_v2_verify_full.txt
—Gold verification: reports/gold_pm_decision_news_verify.txt
—Article group counts: reports/article_group_counts.csv
—Topic coverage: reports/topic_coverage_summary.md
—License/provenance audit: reports/license_provenance_audit.md

License And Sharing

The dataset combines sources with different terms. Use license_class and redistribution_status before sharing rows externally. Public-domain official macro/regulatory rows are the safest shareable content. Non-commercial or unknown-license article text should be treated as internal until reviewed.

Citation

bibtex

@dataset{fenrix_financial_news_lake,
  author = {FENRIX},
  title = {FENRIX Financial News Lake},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\\url{https://huggingface.co/datasets/Scottswi/fenrix-financial-news-lake}},
  note = {Topic-organized PM-decision financial news lake}
}