by Demival Vasques Filho
I just returned from Japan, where I attended the NetSciX (the Winter edition of NetSci) conference. It was an enjoyable event, with inspiring presentations and discussions. Here, I will share with you some of my perceptions about it!
It seems that the humanities still have a low presence in conferences on network science. I was the only person from a History (and humanities) department/institution to raise my hand when a quick and informal poll took place at the conference dinner. Not surprisingly, most attendees came from physics, mathematics, and computer science departments.
Once again, NetSci had a great variety of subjects, exemplified by the keynote and invited talks. Among them, I found interesting the claim by Katy Börner that it is already time to think of data visualization literacy (data literacy by itself is not enough anymore!), and the art of creating visualizations and interpreting complex data.
Network-focused was the talk by Joe Labianca on negative ties and brokers. It was an amusing presentation about how it is more difficult for us to perceive negative ties than positive ones (in the context of animosity and friendship either at work or in our personal lives), and how brokers are rewarded for being original in some environments, by recombining knowledge or bringing new information from other places.
On another note, I was pleased to see the increase in the number (yet small) in works on higher-order networks, that is, interactions between individuals and collectivities. My most recent work at the IEG, that I also presented at this conference, advocates in favor of it, especially when it comes to social networks. Modeling higher-order networks is possible using three methods (so far): simplicial complexes, hypergraphs, and bipartite graphs. My preference lies in making use of the latter. I have been studying bipartite and projected networks for a while now, and in this last study of ours, we have shown that we can think of every social network as having a bipartite structure, as previous works have conjectured. The bipartite network represents the interactions between individuals and collectivities, while the projection onto individuals represents the pairwise interactions that result from the former.
But back to the conference, and its highlights. The use of Twitter data seems to be more and more on the rise. Several studies on different topics, such as opinion dynamics, collective behavior, and fake news, are relying on them. Traditional studies on network structure and network models were still a hot topic in the conference, but both lagging behind dynamics on networks. If feels like social contagion, social behavior, cultural adaptation, and epidemics are the leading interest of network scientists, particularly the latter. Talking about epidemics, it becomes easy to see the importance of such studies on networks, as the current outbreak of the coronavirus gets worse.
Finally, there are a few conferences on networks coming to Europe this year. In June, we have Sunbelt (the international conference on social network analysis) in Paris and the Historical Network Research Conference in Luxembourg, where I’ll give a workshop on bipartite (two-mode) networks. In July, there is NetSci in Rome. Several opportunities to become more involved with networks!