Dataset · Brianferrell787
Financial and general news, 57m rows
Brianferrell787/financial-news-multisource
A unified corpus of 24 public news datasets: 57.1 million rows with English dates and text.
What's inside
An English news text corpus: 57.1 million rows merged from 24 public datasets. The data adheres to a minimal schema of date, text, and extra_fields, organized into parquet shards for streaming and querying via DuckDB. The card also mentions trading date policies.
What you can do with it
- Training a financial news classifier
- Retrieval/RAG for a market assistant
- Fine-tuning LLMs on a news corpus
- Tasks
- Text classification, Text retrieval, Other
- Languages
- English
- Size
- 10M–100M rows
- Downloads this month
- 1.6K
- Likes
- 123
- Access
- on request on Hugging Face
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("Brianferrell787/financial-news-multisource")