Moritz Stephaner and Christopher Warnow used data from SciVerse Scopus (>94,000 publications in total) to map the collaborations of Max-Planck Researchers (inside and outside the MPG) over the last decade. A nice combination of (social) network logic and geography!
The network view shows the Max Planck Institutes and their connections. The size of the circles represents the number of scientific publications for each institute, and the width of the connecting lines the number of jointly published papers between two institutes.
As this was developed in Java, let me for completeness and future reference reproduce the list of (mainly processing related) tools/libraries that went into it:
- processing – an open source programming language and environment for creating images, animations, and interactions
- gephi – an interactive visualization and exploration platform for all kinds of networks and complex systems, dynamic and hierarchical graphs
- JUNG – the Java Universal Network/Graph Framework
- Ani – an animation library for Processing
- GLGraphics – extending the capabilities of the OPENGL renderer in Processing
- traer physics – a simple particle system physics engine for processing
- unfolding – a processing/Java library to create interactive maps and geovisualisations
- tile mill – an open source design studio to create interactive maps
- OpenStreetMap – a free editable map of the whole world.
Well, some items on the list have been covered around here before, the other half seems worthwhile having a deeper look into. The current stable version of processing is 1.5.1, but pre-releases of the upcoming 2.0 version (containing significant changes) are already available. So much to check out, so little time.