Microsoft Academic Graph (MAG) Analytics

Established: April 20, 2018




The Microsoft Academic Graph (MAG) is a knowledge graph of scholarly publications structured around the following entity types: publication, author, author affiliation (institution), publication venue (journals and conferences), field of study (topic). It contains publication dates, as well as citation pairs and of, course, co-authorship data. Because MAG is a knowledge graph, it facilitates powerful scientometric analyses of publication output, impact, collaboration, and more.

What are examples of analyses powered by MAG?

This blog post we authored provides examples of in-depth analytic insights about a scholarly conference, WWW. Please check this page for more examples in the near future.

How can I run my own analytics?

To run your own analytics, follow these steps:

  1. Access MAG
  2. Compute

How can I access MAG?

We provide two ways to access the Microsoft Academic Graph:

  1. The Microsoft Academic Knowledge API available through Microsoft Cognitive Services is free to use and enables you to experiment with the data.
  2. Get Microsoft Academic Graph on Azure storage

How can I learn about how to use MAG data for analytics?

Please take a look at our repository of sample code on GitHub. We hope it helps you get started with MAG analytics! Let us know if the tutorials are helpful, or what type of help you would need to be better able to use MAG data.

To help you learn more about how to work with knowledge graphs and practice on a subset of MAG, we are working on an online course which will be available for free in the near future.

We hope you find MAG useful and we look forward to feedback about what analyses you ran with MAG. Please stay in touch!

Microsoft Academic blog

Microsoft Academic on Twitter

Microsoft Academic Graph project 

Microsoft Academic website (semantic search for scholarly publications)

Microsoft Academic project