Skip to content

Dataset · openalphalab

Multilingual news texts from GDELT

openalphalab/gdelt-news

Corpus of news articles in various languages reconstructed from GDELT for search, generation, and text classification.

What's inside

Contains multilingual news texts reconstructed from the GDELT Web News NGrams 3.0 database (types 1 and 2). Data is stored in parquet format with Zstandard compression and includes dates, language, source links, and administrative metadata. Information about authors, publishers, and countries is missing.

What you can do with it

  • Training and fine-tuning models for news classification or generation
  • Building text-retrieval systems for news archives
  • Analyzing multilingual news streams and trends over time
Tasks
Text retrieval, Text generation, Text classification
Downloads this month
35.8K
Likes
1

How to load it

from datasets import load_dataset

ds = load_dataset("openalphalab/gdelt-news")