This is a copy of the Microsoft Academic Graph corpus of scholarly publications and citations, based on crawls from the open web. Metadata (authors, DOI numbers, journals, citations, keywords, affiliations, etc) is included for more than 125 million publications.
The corpus is a single 27GB zipfile that extracts into about 96GB of flat tab-separated text files, cross-referenced using identifier columns. Schema information can be found in the `readme.txt` file, and usage restrictions can be found in the `license.txt` file.
This corpus downloaded via bittorrent from: http://academictorrents.com/details/1e0a00b9c606cf87c03e676f75929463c7756fb5
The original source was from Microsoft directly: https://academicgraph.blob.core.windows.net/graph/index.html