AFS Sociogram
Copyright (c) 2013, Rinke Hoekstra,
Data2Semantics, VU University Amsterdam & University of Amsterdam
Introduction
This page visualizes the results of an experimental sociogram for the new organization of the AFS.
The experiment consisted of two questions to all tenure and tenure-track staff of the computer science departments of UvA and VU:
- Which 3 people have topics that are currently most interesting for your own current interests? Include at least 1 person from the other university
- Which 3 people have future topics that are most interesting for your future interests? Include at least 1 person from the other university
We use Mike Bostock's D3.js library and the Python network analysis library NetworkX to analyse and visualize the graphs resulting from both questions:
- People are colored according to the louvain method for community detection algorithm [1], implemented in Python.
- The chord diagram distinguishes between incoming links and outgoing links by always using the color of the originating node. Hover over persons to highlight the people they are connected to.
- The force directed graph visualization uses the same coloring, and nodes are sized according to their degree (total number of incoming and outgoing links)
- Mutual collaboration counts double: edges that have an inverse have a weight of 2, all other edges have a weight of 1.
Many thanks to all participants of the CS departments in both universities for providing their data.
Tip: Make your browser window very wide to see the chord and graph side-by-side.
[1] Fast unfolding of communities in large networks, Vincent D Blondel, Jean-Loup Guillaume, Renaud Lambiotte, Renaud Lefebvre, Journal of Statistical Mechanics: Theory and Experiment 2008(10), P10008 (12pp)
Preliminary Conclusions
It seems that the results of the "future" question are a bit biased by optimism. People are very keen to collaborate across boundaries of disciplines in Computer Science, leading to a distorted picture.