AI Co-authorship Network Visualized

Inspiration

Collaboration in AI between the industry and academia has become increasingly important, given that many state-of-the-art AI applications are proposed and implemented by industry-backed entities (think about OpenAI being funded by Microsoft). As a researcher, I want to better understand how those collaborations happened, and how have the collaboration between the industry and academia been, thus giving us insight into how we should look at such collaboration in the future.

What it does

Network visualization of AI conference publication co-authorship between 2017 and 2020

How we built it

First, I retrieved publication metadata from Microsoft Academic Graph using Python and Apache Spark. I used text-filtering and database matching (with Research Organization Registry) to identify researchers who are in the industry. Then, I load and render the network using Gephi and then export the network to a Sigma.js template, then host the visualization using GitHub pages.