Michael Levy of the Center for Environmental Policy and Behavior at UC Davis created a bipartite network of using his coworkers and their preferred journals to illustrate the functional clusters within the highly interdisciplinary lab. He then converted the visualization into a single mode network using ggnet – a ggplot implementation (via the GGally package) and calculated degree, betweenness, and eigenvector centrality for each journal for a more detailed picture of the overlapping interests within his community. He provides his r code for anyone who wants to apply try the excersize with their own lab.
…the campaign [created] a single massive system that could merge the information collected from pollsters, fundraisers, field workers and consumer databases as well as social-media and mobile contacts with the main Democratic voter files in the swing states. The new megafile didn’t just tell the campaign how to find voters and get their attention; it also allowed the number crunchers to run tests predicting which types of people would be persuaded by certain kinds of appeals. Read more
SONIC lab PhD researcher Alina Lungeanu presented a poster titled “The Effects of Diversity on Collaborative Innovative Networks: The Case of the Oncofertility Scientific Subfield” at the 25th Organizational Communication Mini Conference hosted by the University of Oklahoma on October 6, 2012.
Given that an adjacency matrix is a natural representation of a network, a natural visualization is to show the matrix! Forget the tinker toys, who needs nodes and arcs? Amazing retro network visualization demonstrated using an old war horse sample network and d3.js from Mike Bostock. Check it!
…as Facebook has grown over the years, representing an ever larger fraction of the global population, it has become steadily more connected. The average distance in 2008 was 5.28 hops, while now it is 4.74.
What’s all this about p*/ERGM? So you’ve just experienced Prof. Noshir Contractor’s keynote, and he was all over this new-fangled technique for the statistical modeling of social networks, and you’ve never heard of it. Curious? Try one of these articles from our collaborators, generally acknowledged as excellent starting points:
Read both – a slight edge to Robins, Pattison, Kalish and Lusher on the strength of their Figures which may elicit a genuine “Ah, ha!” moment, while the worked examples in Anderson, Wasserman, and Crouch are more substantial.
On Friday, October 21, from 2:00-3:00 p.m. University of Oxford Research Fellow, Bernie Hogan will be giving a presentation in Room 1-483 of Frances Searle Building on the Northwestern University Evanston Campus. This presentation will provide an overview of several studies that explore the phenomena related to how social networks mirror offline networks, albeit not perfectly.
Bernie Hogan is a research fellow at the University of Oxford’s Internet Institute. His work focuses on online identity via real names and pseudonyms. He has published methods for analyzing networks and names in City and Communication, Communication & Society, Fields Methods and elsewhere. His tool for downloading Facebook networks (namegenweb) is used worldwide. His 2009 dissertation under Barry Wellman at the University of Toronto won best Dissertation from ICA’s Communivation and Technology Section.
Yuval Kalish, an assistant professor in the Department of Management of Tel Aviv University will be leading a workshop on Wednesday, September 14 from 2:00-5:30 pm in Frances Searle, Room 1.459.
This workshop provides a hands-on tutorial on how to fit Exponential Random Graph (ERG) Models for social selection using Pnet.
If you plan on attending, please RSVP to Marilyn Logan by 5pm September 13.
ERG models have been referred to as the most promising technique for the modeling of social networks (Snijders, 2007), and has wide applications in the area of organizational studies and communication studies. Topics include: the logic of ERG models, Parameter selection and estimation, parameter interpretation, Goodness of Fit, and troubleshooting convergence issues.
We will discuss Multivariate ERG models if time permits and there is participant interest. Participants are requested to bring their laptops after they have downloaded pnet from: www.sna.unimelb.edu.au/pnet/pnet.html and made sure that it works on their computer.