Skip to content

Dataset · Mearman

OpenAlex Scientific Database Mirror (250+M Works)

Mearman/OpenAlex

Huge multilingual archive of metadata on scientific papers and authors for feature extraction.

What's inside

This is a mirror of the OpenAlex open catalog of scientific publications containing from 10 billion to 100 billion records, including over 250 million scientific works and 100 million authors. The data is presented in text and tabular form in English and other languages, broken down into subsets by entity types, and includes source files in .jsonl.gz format. The copy is transferred from a public S3 bucket with deduplication.

What you can do with it

  • Training graph neural networks and scientific paper recommendation systems
  • Feature extraction and citation analysis in scientometrics
  • Building search engines and RAG databases using scientific metadata
Tasks
Tabular classification, Embeddings
Languages
English, Multilingual
Size
10B–100B rows
Downloads this month
45.1K
Likes
5

How to load it

from datasets import load_dataset

ds = load_dataset("Mearman/OpenAlex")