# Excellence Networks

_Visualizing scientific collaboration_

Published: 2015-09-07

Live project: [http://excellence-networks.net/](http://excellence-networks.net/)

This web application visualizes how successful research institutions collaborate. How does co-authorship with others affect the probability that publication get cited?

## Key visuals

- Collaboration network centered on University College London, linking to Oxford, Cambridge and Max Planck
- Biochemistry and molecular biology network of 800 institutions, average best paper rate 35.1%
- Network radiating from the U.S. Food and Drug Administration to NIH, universities and health agencies
- Chemistry collaboration network centered on Max Planck Gesellschaft, best paper rate 26.2%
- Start screen with four view previews and buttons to select a subject area
- World map of IEEE USA engineering collaborations, institutions colored by citation performance
- World map of 402 computer science institutions, average best paper rate 20.8%
- Geographic layout with institutions colored by country, clustering the USA, China, Japan and Europe
- Ranked table of materials science institutions by country, publication count and best paper rate
- Stanford University network, tooltip shows best paper rate rising to 65.6% cooperating with Peking University
- Collaboration network centered on the Russian Academy of Sciences, dominated by low best paper rates in red

![Collaboration network centered on University College London, linking to Oxford, Cambridge and Max Planck]()

The application is based on papers (articles, reviews and conference papers) published between 2007 and 2011 in 17 subject areas. For each subject area, those institutions (universities and research-focused institutions worldwide) were selected in the SCImago Institutions Ranking which published at least 500 papers in the publication period. We then estimated statistical models (Bayesian multilevel logistic regressions) for the 17 subject areas. The statistically estimated best paper rates which the reference institutions have achieved with its network institutions are visualized. The best paper rate gives the proportion of highly cited papers from an institution and is considered generally as a robust indicator for measuring citation impact. Co-authorship networks (based on institutional affiliations) show how successfully overall an institution (reference institution) has collaborated compared to all the other institutions in a subject area, and with which other institutions (network institutions) an institution has collaborated best.

![Institution bubbles colored by best paper rate, red to blue, with subject-area average 19.2%]()

Insitutions are colored by the probability that their papers are cited by others. The application allows to investigate if and how that probability changes when specific institutions collaborate.

![Stanford University network with tooltip showing best paper rate rising to 65.6% in cooperation with Peking University]()

## Credits

Based on research from Lutz Bornmann, Rüdiger Mutz, Moritz Stefaner, Felix de Moya Anegón. Thanks to [Lynn Cherny](http://ghostweather.com/) for support with the web development and network analysis!

---
[View on truth-and-beauty.net](https://truth-and-beauty.net/projects/mpg-excellence-networks)