Here is Dr. Cody Dunne speaking about a new information visualization technique called “Network Motif Simplification” at the recent CHI 2013 conference in Paris, France.
Networks, no matter how complex, are composed of simpler, smaller structures, called motifs. Some of these structures are easy to identify, like the pattern of a “star” where a single node acts as the sole connection to a connected component for one or more “pendant” nodes with a single tie. Another common pattern are nodes that are “parallel bridges” which share the only two connections they have with two or more other nodes. These common structures can be identified and removed and replaced with more efficient and comprehensible representations.
The result is a simplification of the network visualization, removing clutter to reveal the core structural properties of interest.
A complex network of voting relationships in the
2007 United State Senate is reduced to a simplified form
This method for collapsing complex network graphs into simpler forms has been implemented in NodeXL. Look for the feature in the NodeXL Ribbon menu, in the NodeXL > Analysis > Groups > Group by Motif… option.
NodeXL implements network motif simplification
The feature allows users to select the types of motifs that should be recognized and collapsed:
Here are recent graphs of Twitter networks for several news media outlets :
@cnn
bbcworld
@washingtonpost
theeconomist
@latimes
@FT OR @financialtimes
@nytimes
cbcnews
@newshour
nbcnews
bbcnews
cbsnews
The common “broadcast” structure is common to most of these news media outlets, it appears as a “hub and spoke” pattern. The people at the end of these spokes are the “audience”. Some of these news networks have many more “isolates” or “brand” mentioners – these are the grids of individuals with no connections to others. In contrast some contributors are densely connected in communities of discussion formed around various topics.
SF Online Community MeetUp is the free monthly gathering of online community managers, enthusiasts, and innovators to meet and discuss tools and strategies for building and managing effective communities.
During our March 26 Meetup we’re happy to welcome Marc A. Smith, Chief Social Scientist at Connected Action Consulting Group for his talk, “Charting Collections of Connections in Social Media: Creating Maps and Measures with NodeXL.” The talk will explore how the Social Media Research Foundation, an organization formed to develop open tools and data sets and to foster scholarship related to social media, is using NodeXL to create social network maps. Learn how you can use this free and open tool to map public social media conversations happening among your online community across social networks. Find out how NodeXL can augment your existing community management practices to identify key influencers in your community, discover relationships and strategic hashtags, and more.
I am delighted to be attending and presenting a workshop on social media network analysis at the University of Kentucky’sLINKS 2013 program on June 7th, 2013.
This week long program has for many years provided intensive training in network methods, research, and tools.
I am excited to attend some of the program and meet researchers and students working on networks of all sorts. I will do a short hands-on talk about NodeXL and a longer day devoted to the broader ways networks are useful for the study of social media.
Hidden within social media streams are structures that identify the most influential voices on any topic. Social network analysis and visualization can take millions of messages and reveal the shape of the crowd and the people at the center of it. Using the free and open NodeXL application, this talk demonstrates the tools and methods needed to create detailed maps of any social media topic. Learn to map and analyze social networks extracted from email, Facebook, Twitter, YouTube, message boards, and the WWW. No coding or prior experience needed!
Workshop: Tuesday, April 16, 2013 From: 6:30pm – 9:30pm
Intended Audience: Social media managers and analysts, marketers, collaboration and enterprise IT, advertisers, event planners, journalists,
Knowledge Level: all skill levels, beginners particularly welcome. Should have an interest in social media. Any experience with a spreadsheet is a plus!
Workshop Description
Social media conversations are clumpy. People tend to follow and reply to people who share their views so distinct clusters emerge in many social media discussions. Often these sub-groups have distinct ways of using language, point to different URLs, and mention different hashtags, even when talking about the same topic. Simple, free and open tools can now collect and analyze these clusters of discussion, highlighting the contrasting themes in the conversation. Learn how to perform key tasks like:
I willspeakat the 2013PASS Business Analytics Conferencein Chicago on April 12, 2013. The event focuses on data discovery, data exploration and visualization, predictive analytics, content management and architecture, collaboration, information strategies, and more. I will speak about social mediaData Analytics and Visualization in a presentation titled: Charting Collections of Social Media Connections with NodeXL.
This is a sample NodeXL graph that represents a network of 106 Twitter users whose recent tweets contained “passbac”. The network was obtained on Wednesday, 30 January 2013 at 01:09 UTC. There is an edge for each follows relationship. There is an edge for each “replies-to” relationship in a tweet. There is an edge for each “mentions” relationship in a tweet. There is a self-loop edge for each tweet that is not a “replies-to” or “mentions”. The tweets were made over the 7-day, 5-hour, 32-minute period from Tuesday, 22 January 2013 at 17:40 UTC to Tuesday, 29 January 2013 at 23:12 UTC.
Learn to make your own network maps of social media at PASSBAC 2013!
This April 8 and 9, 2013 an NSF funded workshop called Kredible.Net to be held at Purdue University will bring together researchers studying reputation and social roles in social media.
The grant will help researchers investigate how social media, especially Wikipedia articles and editors, shape public knowledge. The project aims to build a research community and to propose a research agenda for the study of reputation and authority in informal knowledge markets, such as Wikipedia.
This study integrates network and content analyses to examine exposure to cross-ideological political views on Twitter. We mapped the Twitter networks of 10 controversial political topics, discovered clusters – subgroups of highly self-connected users – and coded messages and links in them for political orientation. We found that Twitter users are unlikely to be exposed to cross-ideological content from the clusters of users they followed, as these were usually politically homogeneous. Links pointed at grassroots web pages (e.g.: blogs) more frequently than traditional media websites. Liberal messages, however, were more likely to link to traditional media. Last, we found that more specific topics of controversy had both conservative and liberal clusters, while in broader topics, dominant clusters reflected conservative sentiment.
When a particular node in a network is of special interest it can be useful to create a network visualization in which it is located at the center of concentric rings of vertices.
NodeXL supports a “Polar” layout in which each vertex has two values that govern its location: distance from center (“Vertex Polar R”) and the angle around the clock (“Vertex Polar Angle”).
Using a random network, we added two columns to the Vertices worksheet that we called: Ring (or “Vertex Polar R”) and Rotation (or “Vertex Polar Angle”). We then assigned values for the “Ring” and the “Rotation” for each Vertex:
These values can then be mapped to the location for each Vertex using the NodeXL Autofill columns feature:
When these values are applied to the network visualization and the layout is set to “Polar” the visualization repositions each vertex into a position around a ring. The values are set by mapping Vertex Polar R to “Ring” and Vertex Polar Angle to “Rotation” and then selecting “Autofill”. The result is a single ring plotted around a core:
To see this more clearly, I built a larger random network with 100 vertices and added two more “rings”. The resulting image looks like this:
Dr. Cody Dunne at CHI 2013 (Photo Credit: Ben Shneiderman)
