Complexity and Social Networks Blog

« January 2007 | Main | March 2007 »

28 February 2007

WYSIYN - Build your own social networks

It was only a matter of time. Remember those years when the creation websites was only something for programmers? But then came WYSIWYG HTML editors and thereafter blogs and Wikis, which allowed almost anyone to participate in the web. However, while many of us joined online social networks, the creation was the domain of software developers of the respective platforms (i.e. MySpace, Xing, ASW , facebook to name a few). This is about to change with "what you see is your network" (WYSIYN) Web20 solutions such as ning. There are other companies like coghead or teqlo which allow users to take advantage of other web technologies in an intuitive way. Ning, of course in endless beta, offers the usual features (see a video walkthrough and another review) most of us are now used to from the existing online SN platforms: public/private network, member profiles, forums, blogs, picture and video sharing.

This development raises an interesting set of questions. More on the business side: Will the private online social networks (OSN) canibalize the business models of the existing "professional" platforms? Will this lead to an uncontrollable rise of thousands of OSN that are short lived (Sustainability of OSN was adressed earlier by Ines) just like many blogs (Managing an online community costs certainly more time and effort than running a blog). There are still millions of people who are not using online social networks at all. On the other hand, should these very specific communities be of interest to marketing, advertising, campaign managers anytime soon? Will the majority of individuals create the OSN to manage their offline social networks or would they rather create these OSN around an area of personal interest. The latter might support the notion that we will see the emergence of short-term OSN in the near future. I am also wondering what kind of innovative ways these individuals might develop to market their OSN and keep it alive facing competition from big players, the usual time constraints of daily life and decreasing motivation to participate.

More on the network side. Yesterday Thomas underlined the importance of having critical mass of users in online communities which consists of members that can be grouped around certain roles/ characteristics. What is the maximum number of OSN people are willing to connect? Would Meta-SN Plattforms be next once user demand spurred by their multiple memberships in OSN force even the big platforms to agree on some joint standards (i.e. OpenID) to be more user friendly, thus making it easier to manage multiple online identities and prevent entering redundant information over and over. The big players are already opening up to third-party providers according to a recent article in BusinessWeek.

Finally, would the readership of this blog be intersted in an OSN of "Social Network and Complexity Researchers" instead of our current attempt of creating a Global Social Networks Researchers Mash-up which so far only attracted 43 individuals to join.

Posted by Alexander Schellong at 12:00 AM

27 February 2007

Show me your links and I tell you about your political ideology - Applying Network Theory to the War Blogosphere

While doing research on a different topic I stumbled over an article in the Journal of Computer-Mediated Communication by Tremayne, Zheng, Kook Lee and Jeong on "Issue Publics on the Web: Applying Network Theory to the War Blogosphere". They looked at 79 political blogs and 899 posts about Iraq before and during the year of 2003 to examine the process by which certain blogs become hubs of activity while others remain on the periphery. In addition they tried to analyze the role of political ideology in network formation.

One of the many findings presented was that...

"blogs coded as very conservative were twice as likely to link to other blogs as were blogs coded as liberal or very liberal. Conversely, blogs coded as very liberal were more than twice as likely to link to media sites as blogs coded as conservative or somewhat conservative. These results lend support to the notion that conservatives distrust the mainstream media more than liberals do.(...) While the New York Times and Washington Post were the most linked-to media overall, a finding also reported by Adamic and Glance (2005), liberal bloggers were more likely to link to these media than were conservative bloggers."

...which probably confirmed what many of you expected. Next time you read a blog make sure you take a closer look at the links to determine your information on the political spectrum. Going further, an analysis of the links of an individuals blog would allow one to learn more about the individuals information sources which combined with other information easily obtainable on the net should be of interest for campaigners and marketing experts to influence or spread information.

Posted by Alexander Schellong at 12:08 AM

26 February 2007

Critical Mass in User Communities

A critical mass of users who actively engage in information exchange and knowledge sharing activities is crucial to keep an online community (of users) up and running over time. But how can critical mass in online communities be established and sustained over time? What are promising strateg(ies) for community operators to turn online communities into places that are worthwhile lingering?

One potential strategy can be observed in aSmallWorld.net which is a prominent online community. Critical mass was established by bringing together two types of people:

(1) celebrity-type people, and
(2) internet "aficionados" who love spending their time on the web and in interactive chat rooms.

While the celebrity-type people make sure that the site receives significant media and the general public's attention, internet "aficionados" generate the traffic which is important to make community look active.

Any other ideas how the creation of critical mass can be established?

Posted by Thomas Langenberg at 5:14 AM

23 February 2007

gaming google

One of those interesting phenomena that I have noticed is the gaming of the google algorithm (at its foundation, a network-based algorithm utilizing various measures of network centrality) through posting of comments to this blog. In the google world we live in, the more links that point to you, the more traffic you get. And more traffic equals money. Given the growing importance of blogs in the web, a number of services have sprung up to place comments on blogs to point to some website. For a while, the comments were easy to pick out-- for example-- "great blog. www.---.com". More recently, however, the text of these comments are more specific to the entry, but often barely intelligible-- which may reflect some natural language automated system, or individuals who are being paid to post quick sentences with links to many blogs. Anyhow, it is both an interesting cat and mouse game on the world wide web, and a bit of an annoyance (a la spam) for us. We are generally catching these, but if one gets by, let us know.

Posted by David Lazer at 10:31 PM

19 February 2007

The scaling laws of human travel

A friend of mine recently pointed me to this fun Nature paper entitled "The scaling laws of human travel". Essentially, these guys took the Where's George? dataset and successfully characterized the movement of $1 bills across the United States.

For those not familiar with the Where's George project, the idea is to record the unique identifier of US dollar bills and stamp them all with the "WheresGeorge.com" logo. The project has set up a web site where people who find a Where's George bill log in and record the bill's unique identifier and the place they found it. There are millions of bills currently circulating freely across the US and beyond. (I saw a $1 bill stamped with "Where's George" in Addis Ababa, Ethiopia last fall!)

While it's an interesting standalone website, the subsequent data generated from this project seems to provide insight into the spatiotemporal dynamics of a (fairly) unbiased random sample of the US population. The authors point out that the distribution of the distance travelled decays as a power-law and show the trajectories of individual bills can be fit well using a two parameter variation of a continuous-time random walk (CTRW).

Although I would have liked the authors to discuss the implications associated with human traveled being governed by a CTRW rather than pure Levy flight dispersal (ie: so what?), their rigorous analysis of this dataset is a great example for other researchers with similar spatiotemporal data. The paper also is a good place to start when brainstorming about what other comparable datasets are out there that could help us get a better handle on the complex dynamics associated with the behavior of our society.

Posted by Nathan Eagle at 9:20 AM

Extending the Technology Enactment Framework - PNG Working Paper

Jane Fountain’s book “Building the Virtual State” introduced social science researchers to the technology enactment framework (TEF). This working paper presents further modifications to the revised TEF by Okumura who introduced key actors that influence technology enactment. I propose a fourth actor group, the citizen and further causal relations between existing actors and the organisational setting. The revisions towards a more hybrid TEF between an actor-centric and institutional approach allows overcoming some of the limitations brought up by the framework's critics such as the absence of socio-technical systems theory.

Posted by Alexander Schellong at 12:15 AM

16 February 2007

Structure versus Content in Network Analysis

There is a fundamental assumption underlying analyses of networks investigating the effects of ties. The assumption is of a commonality or regularity in the presence of ties among dyads and the content of the interactions taking place across these ties.

If the tie relationship is defined as the classical friendship, trust, or advice type ties, this assumption may not be particularly problematic. As more and more network studies construct networks from data harvesting (e.g., email logs, text analysis, etc.), this assumption merits more scrutiny.

Click below for additional discussion with examples.

First, consider a less problematic example - email logs. Here, the network is formed by the collection of asymmetric sender-recipient ties, and no content is coded. Although "flame wars" are a well-known phenomenon of electronic communication, it remains unlikely that the pattern of one's email network for friends is similar to that of one's enemies or even acquaintances. Indeed, VisiblePath's goal is to infer trust relationships from email, IM, and scheduling logs. That is, information about the *content* (or consequences) of ties can be inferred from network structure. This analysis provides novel insights that can be difficult to obtain by other means.

Consider another example, the person-gadget bipartite network of professional reviewers of tech gadgets harvested from the text analysis of their reviews. Absent the content of the review, a network representation of the presence of review ties is unlikely to provide any new insights. The domain of interest or expertise of a reviewer and his/her community of peers could likely be inferred from the network, but these could be found at least as easily by examining the staff rosters across the periodicals used as sources. Absent information on the content of the tie, where the content varies independently from structure, novel insights from exclusively structural
analyses will be more limited.

Now take the example of automated data harvesting of blog entries, for example, where interpersonal networks could be constructed based on author-subject ties (let's limit the subjects in this example to other individuals identified by name in the text of the blog entry). Entries can be about topics an author finds inspiring or infuriating. Their subjects could be just as likely to be praised as pilloried (especially on political blogs). In such a case, then the range of possible inferences about the consequences of these ties is constrained as in the previous example.

In these problematic cases, it is decidedly NOT necessarily true that the network does not matter. Rather, the problem is that divorced of the content of the tie, it can be hard to know whether or how the network matters. Null effects from network analyses could arise from the lack of any effect, or they could indicate that the network definition was insufficiently sensitive to the content of the ties. The solution is to encode some of the content within the network. So for the examples above, including the number of stars as a tie weight for reviewer-gadget ties, or the ratio of positive to negative adjectives in the text around a subject's name in the author-subject ties of blog entries, could provide for more novel insights from a network analysis.

As more network analysis is conducted on data harvested in automated ways, it is important to consider the possible importance of tie content, and how it may be incorporated into structural representations to best support revelatory analyses.

Posted by Brian Rubineau at 12:33 AM

15 February 2007

Mobile phones in the developing world - Part III

I recently came across another interesting news article about the adoption of mobile phones in developing countries. This particular article focuses on mobile phone adoption from the perspective of companies who are selling mobile phones in India. These companies are scrambling to make low-cost phones that will endure dust, heat, and long periods of time without recharging. What strikes me most about this article are the lengths that people living under impoverished conditions will go to connect, and stay connected, with their social networks. Having little else, they are still willing to spend a large amount of their income on a single piece of social technology. Yet, from the social support perspective, this makes perfect sense. If people in India are anything like those studied in America, they exchange different kinds of support with different kinds of ties, and mobile phones enable them to stay connected to a variety of ties like never before. On the other hand, while Americans often use their networks to get ahead, people in India may need them just to get by.

Posted by Jeff Boase at 12:04 AM

14 February 2007

"The Craft of Connection" and a firm's ability to exploit the things they know

The strategy+business magazine recently launched a very interesting article on the "The Craft of Connection" within firms. The authors Tim Laseter and Rob Cross argue that the management of internal social networks is an important task which managers need to address as well when it comes to increasing firm performance.

Along these lines, in a recent theoretical study, Ines Mergel, Christopher Tucci, and me found that centrality of managers as well as the overall density of such firm internal social networks are important variables which explain when and why some firms are more capable of exploiting their knowledge than others. A working copy of the paper is available upon request from the author of this blog entry.

Posted by Thomas Langenberg at 12:00 AM

13 February 2007

How does the way we process information relate to how we search for it?

Some days ago I attended a talk on human information processing by Thomas Mussweiler from the University of Cologne who spoke at the Columbia Business School. Mussweiler and colleagues conducted an impressive number of experiments on the mechanisms and influences of individual information processing. A simple example would be to ask you to determine your best athletic performance. You have two basic options: 1) You think of every single athletic moment in your life, i.e. you engage in absolute information processing, or 2) you compare what you recollect as some of your best performances to a given standard, e.g. a famous athlete’s performance (or a famous couch potato’s performance). Not surprisingly it turns out that comparison allows to process information in a more efficient manner.

Mussweiler went on to talk about various factors that influence the comparisons we make, most importantly the standards we employ for comparing information. His experiments used a technique called “priming” to activate certain standards – for example, subjects were asked to judge a trait in a person. The result shows that priming a trait concept (such as aggressiveness) will induce the subject to judge the target person according to that trait. In other words, once activated, standards are spontaneously compared to the target person.

While I was listening to the talk, I kept asking myself how the way we process information relates to how we search for it. Some possible bridges might be that the search itself is the result of some form of information comparison (my search is triggered by a comparison of the information I have to a “standard”, which is the knowledge I believe I need to possess), and/or that we subconsciously use standards to determine the source to turn to when searching for information. I don’t know if there’s literature out there that links cognitive psychology to advice networks, but it's definitely something useful to look into.

Posted by Maria Binz-Scharf at 12:48 PM

12 February 2007

"networking" for a cause

The proliferation of "social networking" functionality online is striking. A large number of websites, whose primary objective is clearly to support some other goal, now have features that allow "members" to post material and indicate their relationships with others in the community. (This is rather different than more general social networking sites, such as facebook, myspace, etc, whose business models are to get traffic and data on individuals and somehow monetize that.) Presumably the intuition here is that this increases the attachment of individuals to the cause or the product, and helps facilitate (in particular cases) collective action. It would be interesting to rigorously assess whether these features actually have any impact. While I see the logic, I will admit to some scepticism.

Some examples:

The Barack Obama campaign

United Methodists

Toyota hybrid owners

Posted by David Lazer at 10:20 PM

6 February 2007

Strong ties are those that look for you…

See excerpts on New York Times story on disappearance of Microsoft researcher James Gray. This is a remarkable illustration of social capital—the social capital story underpinnings are: (a) that Gray had many ties; (b) these ties were to people who control important resources; (c) that these relationships were strong enough to mobilize these individuals; (d) that their network of relationships enabled effective collective action, and (e) the role that technology has played in allowing a bottom up, distributed, effort in the search.

On this last point, check out/ or even help look for James Gray, by going to the Amazon Mechanical Turk. AMT is a tool to allow distributed work on very large projects. From the website:

“During the last 5 days, Mechanical Turk workers looked at more than 560,000 images from 3 satellites, covering nearly 3,500 square miles of ocean. A group of experts is currently reviewing the images that workers identified, and sending their results to the appropriate authorities.”

My best wishes to Dr. Gray and, to his friends and family in their efforts to find him.

DL

____________________________________

To see full NYT article .

February 3, 2007
Silicon Valley’s High-Tech Hunt for Colleague
By KATIE HAFNER
SAN FRANCISCO, Feb. 2 — When James Gray failed to return home from a sailing trip on Sunday night, Silicon Valley’s best and brightest went out to help find him. After all, Dr. Gray, 63, a Microsoft researcher, is one of their own. The United States Coast Guard, which started a search Sunday night, suspended it on Thursday, after sending aircraft and boats to scour 132,000 square miles of ocean, stretching from the Channel Islands in Southern California to the Oregon border...

Teams turned up nothing, not so much as a shard of aluminum hull or a swatch of sail from Dr. Gray’s 40-foot sailboat, Tenacious.
In the meantime, as word swept through the high-technology community, dozens of Dr. Gray’s colleagues, friends and former students began banding together on Monday to supplement the Coast Guard’s efforts with the tool they know best: computer technology.
The flurry of activity, which began in earnest on Tuesday, escalated as the days and nights passed. A veritable Who’s Who of computer scientists from Google, Amazon, Microsoft, NASA and universities across the country spent sleepless nights writing ad hoc software, creating a blog and reconfiguring satellite images so that dozens of volunteers could pore over them, searching for a speck of red hull and white deck among a sea of gray pixels.
…
On Tuesday evening, as the Coast Guard’s search continued, Joseph M. Hellerstein, a computer science professor at the University of California, Berkeley, sent out an e-mail message with the subject: “Urgent ... Jim Gray.” One recipient, Sergey Brin, co-founder of Google, wrote back within an hour, and offered to enlist Google Earth’s satellite imaging expertise.
By Wednesday, Professor Hellerstein had started a blog and earth sciences experts at the Ames Research Center of NASA in Moffett Field, Calif., had sprung into action. They secured the promise of help from a high-altitude aircraft equipped with a high-resolution digital camera that was already scheduled for a flight Friday from Dryden Research Center in Southern California but whose pilot could make sure his path included the search area.
…
“The number of people who feel they owe him in so many ways, personally and professionally, as a role model and friend is incredible,” Professor Hellerstein said.
Dr. Gray is a leader in the field of database systems and transaction processing and has received several computer science awards, including the prestigious Turing Award in 1998.
And there is an infinitesimal degree of separation between Dr. Gray and nearly everyone involved in the search for him.
“Nearly every major research project he worked on has been hugely influential on later research and products,” said Phil Bernstein, a principal researcher at Microsoft who is a colleague of Dr. Gray.

Posted by David Lazer at 8:35 PM

5 February 2007

Cross boundary collaboration and eGovernment: PNG Working Paper

Administrative and political leadership need to use their growing understanding of
eGovernment to come up with strategies that help them crossing the boundary between organizatonal units for better collaboration and coordination. The PNG working paper "Crossing the boundary - Why putting the e in Government is the easy part" reviews the current status of eGovernment projects and research from around the globe and offers additional insights in how to overcome these challenges.

Figure: Modified Gartner eGovernment Hype Cycle

Posted by Alexander Schellong at 12:00 AM

3 February 2007

"Visualizing Network Dynamics" competition

A quick announcement of the "Visualizing Network Dynamics competition", organized by Katy Börner, Stephen Uzzo, Marcia Rudy, Elisha Hardy (truth in advertising, I am one of the judges).

Below are excerpts from the announcement.

The "Visualizing Network Dynamics" competition will be held as an integral part of NetSci07. The competition will invite researchers, practitioners, and educators from such diverse disciplines as anthropology, sociology, history, social psychology, political science, human geography, biology, economics, communications science but also art and design to submit the best-of visualizations of evolving networks, activity patterns over networks or combinations of the two. Competition applications will comprise large resolution static images or video footage together with a detailed explanation of datasets used, analysis or modeling techniques applied, and visualization design. Applicants will also be asked to list and explain major insights gained and to discuss the value the visualization might have for educational purposes.

The competition aims to harvest the best examples of meaningful (as opposed to purely artistic) network dynamics visualizations, to raise the bar for the documentation and communication of the process applied to generate those visualizations, and to sensitize people to the importance of visualization for formal and informal education and the communication of science in general.

Correspondingly, visualizations will be judged based on:

* Truthfulness of the data representation
* How well the visualization serves the needs of its ‘clients’
* Quality of data preparation and analysis
* Layout and design of the visualization
* The significance of insights gained
* Educational value
* Visual appeal
* Description of work

Awards

The Top-3 winning entries will get free registration to the NetSci Conference 2007 and cash prices in amounts of $100, $200, and $300 that are sponsored by the Cyberinfrastructure for Network Science Center at Indiana University.

The Top-25 winning competition entries will be printed in large format and displayed at the Network Science Workshop and Conference. Winning animations will be projected in large format.

A DVD with all valid entries and their accompanying information will be shared with all Conference attendees. All valuable entries will also be made available online as a general, free resource for anybody interested in the study or communication of dynamic networks.

Posted by David Lazer at 2:32 PM

2 February 2007

Government Social Software - SNS in Japan Part IV: Connecting the cases to the literature

Today's post ends my entry series on the use of local Social Networking Services by Japanese municipal governments. If you would like to read the entries in order just click here: Introduction, Yatsushiro Case, / Nagaoka case, discussion (organizational aspects).

Even without knowing the respective research and terms interviewees made the correct assumptions about social networks or tell stories reflecting results of social networks and social support in disaster literature. Drawing for example on the narrative of the family that was helped by many strangers after a the mother of a sons friend (weak tie) wrote about their flooded house in her “Gorotto Yatchiro” blog which supports Granovetter’s weak tie and Burt’s structural hole role in non routine activities (2004; 1983). Those interviewees who joined the local SNS found new friends on the platform and expanded their social network as concluded by Tindall and Wellman (2001). Furthermore, Soiga NPO is a great example how an organization, once brought into existence for one set of purpose (environmental activities), can also aid others for different purposes described by Coleman, thus constituting social capital available for use (1988). The NPO’s blogs were considered a trusted source and can provide an alternative to the mass media which is regarded by many individuals as a more credible source of risk information than government (McComas, 2001). A centralized approach to the provision and publication of local information might not be fine-grained enough to cater to the viral and capillary spread of word-of-mouth information anyway. This informal interaction can only be supported by recognizing the peer-to-peer nature of local interaction which is distinct from the conventional many-to-many, few-to-many, or one-to-many broadcast nature of other online interaction (Foth, 2006). In the past this role was taken by neighbourhood organizations which are already impacted by demographic and cultural change (young generations are not really interested in joining).
Finally, if the majority of the population would be represented on local SNS platform, sociograms could provide snapshots of networks and interaction structures. From these types of diagrams government and citizens can visually identify emergent positrons and clusters of interaction. By examining these patterns of mediated and unmediated interaction they could gain an added perspective on communication structures that underpin explicit community processes as well as those that support affective, less instrumental behaviors (Garton, Haythornthwaite, & Wellman, 1997). Privacy might be a concern for citizens of course. At the moment, local SNS can serve the functions of managing and building social networks. In disasters it covers the areas of “observe and report” and “warn and inform”. Along the lines of La Porte, I argue that the design and rules of the network constrain the character, use and content of member roles and exchanges and the network (1996). Consequently, local SNS could support the community and government beyond its current scope.

Sidenote: As I heard this week MIC is planing to extend their local SNS pilot with 10 other cities. I will keep you posted.

Posted by Alexander Schellong at 12:10 AM