All you theoretical superconductor folks certainly seem to be good collaborators. I don’t know about polite, but network analysis seems to indicate you have a pretty tight community.

With roughly 17,000 authors listed in 7,000 March meeting talks, it’s hard to imagine finding any order to it all. But science is a collaborative pursuit, and Jake Hofman thinks network analysis could say something about the various physics communities hidden in that massive list.

Hofman, a graduate student at Columbia University, downloaded the list when the abstracts came online a few months ago. He used the data to create a network, linking physicists who were listed as co-authors on a talk.

Many presenters were only connected with a handful of other physicists. But there was also a group of 6755 physicists (the largest connected component, for network nerds) who were all linked to each other in one way or another – think six degrees of Kevin Bacon. Hofman then ran an community-identifying algorithm he’s working on to look for clusters of physicists that seem particularly closely linked.

To the right you’ll see the visualization he ended up with. It’s a little tricky to interpret (and also quite preliminary), but the basic idea is that each of the 6755 physicists is shown in order on both the x- and y- axis. Blue points indicate co-authors, and white points show no affiliation. Large clumps of blue show communities, most of which seem to correspond to particular subfields.

Hofman just started analyzing the data, but he’s already highlighted some particularly collaborative communities. The visualization also shows that there are multiple communities that focus on the same thing. There are, for instance, at least two fairly large graphene communities. Rivalry!

For the nitty gritty details on Hofman’s approach, check out his paper in the arxiv.

All you theoretical superconductor folks certainly seem to be good collaborators. I don’t know about polite, but network analysis seems to indicate you have a pretty tight community.

With roughly 17,000 authors listed in 7,000 March meeting talks, it’s hard to imagine finding any order to it all. But science is a collaborative pursuit, and Jake Hofman thinks network analysis could say something about the various physics communities hidden in that massive list.

Hofman, a graduate student at Columbia University, downloaded the list when the abstracts came online a few months ago. He used the data to create a network, linking physicists who were listed as co-authors on a talk.

Many presenters were only connected with a handful of other physicists. But there was also a group of 6755 physicists (the largest connected component, for network nerds) who were all linked to each other in one way or another – think six degrees of Kevin Bacon. Hofman then ran an community-identifying algorithm he’s working on to look for clusters of physicists that seem particularly closely linked.

To the right you’ll see the visualization he ended up with. It’s a little tricky to interpret (and also quite preliminary), but the basic idea is that each of the 6755 physicists is shown in order on both the x- and y- axis. Blue points indicate co-authors, and white points show no affiliation. Large clumps of blue show communities, most of which seem to correspond to particular subfields.

Hofman just started analyzing the data, but he’s already highlighted some particularly collaborative communities. The visualization also shows that there are multiple communities that focus on the same thing. There are, for instance, at least two fairly large graphene communities. Rivalry!

For the nitty gritty details on Hofman’s approach, check out his paper in the arxiv.