Knowledge Matters™

Two weeks ago I purchased a new computer, which came with Microsoft® Vista Business loaded. I haven't gotten used to Vista yet and I am not sure I really like it. That said I've always liked Excel, and Excel 2007 seems better than ever. One of the really nice add-ons is Microsoft .NetMap , which installs a template capable of doing some rudimentary network analysis. Consider the diagram below, which is my email traffic, displayed using a Fruchterman-Reingold force-directed spring algorithm.

The add-in analysed Outlook 2007 and identified all 532 unique vertices and the corresponding 994 unique edges that make up my e-mail network. It also identified the density of the network as 0.004, which is quite sparse given 1.000 is the possible score. Consider now the same network displayed in a spiral format. ...

read more ...

I've just finished reading Methodological Pitfalls in Social Network Analysis by Nicholas Marschall. The central theme is that current methods produce questionable results, which is precisely why I read the book.

Running to 86 pages the book is an easy one-sitting read. For what it is it's also expensive. The book is a translation from German so in some places the English is - well unusual. Looking beyond this small problem, it appears to be a student or scientist research justification, or perhaps a short synopsis of a PhD, which means the style is very academic, but it is interesting!

Marschall quite rightly says data collection approaches colour results, and are full of implicit assumptions. He comes to the conclusion that size reduction and transformation processes, which are quite common in published studies, can significantly change the results of an analysis. ...

read more ...

Matthieu Latapy is a network researcher at the National Centre for Scientific Research (Centre National de la Recherche Scientifique) in France. He is doing network research to counter peer-to-peer paedophile content - something I think we would all agree is very worthwhile!

Matthieu is conducting an experiment aimed at understanding how a real-world spreading phenomenon occurs via links between websites. In particular he notes that information diffusion is increasingly orchestrated by bloggers instead of the mainstream media. Matthieu's experiment takes the form of a viral marketing approach - hence the name Happy Flu.

To participate all you need to do is: ...

read more ...

This isn't a blog-post about explosives, although that might be interesting; rather it is a plug for the blogs of two people whose work I admire. It's no secret I am interested in network analysis and both these blogs have almost exclusive network analysis themes. The TNT bit stands for "The Network Thinker ", which is the blog of Dr Valdis Krebs - I wish I could be as creative with names. Connectedness is the blog of Dr Bruce Hoppe.

I particularly like Connectedness because of its exclusive network analysis theme. I just wish Bruce would blog a bit more. He's had a few interesting posts of late. Take a look at this post on centrality . Bruce correctly identifies centrality is important in every network; and lets be honest it's the property we are most fascinated with! I would caveat Bruce's observation by saying be careful with what type of centrality you are talking about. For example, in-degree centrality is about popularity and out-degree centrality is about activity - two very different things I suggest!

read more ...

From time to time I run a half-day seminar called “Introducing Network Analysis as a Research Technique ”, followed by a practical workshop that builds on the mornings activities and introduces participants to UCINET and NetDraw .

Here are the seminar slides . The seminar is aimed at new researchers. Typically the seminar occurs in the morning and is organised into three sessions.

Here are the workshop slides . The workshop is aimed at absolute beginners, and introduces participants to UCINET and NetDraw . It builds on the seminar, although attendance at the seminar is not a prerequisite.

read more ...

On Friday the 4th of July 2008 I am once again running a half-day seminar called “Introducing Network Analysis as a Research Technique ”. This time the seminar will be followed in the afternoon by a practical workshop that builds on the mornings activities and introduces participants to UCINET and NetDraw .

The seminar and workshop are part of Canberra University’s Inter-University Research Workshop Program , and are aimed at new researchers. The workshop is also of use to anyone interested in network analysis, but is aimed at absolute beginners. Both the seminar and workshop are offered free of charge to research students and researchers from all institutions, but priority access will be given to students and staff from participating universities. Costs may apply to other participants. If you are interested and you are in Canberra on Friday the 4th of July 2008 you can register here . Further details about the seminar and workshop are below. ...

read more ...

Every now and then I come across an intriguing article that is also seminal to my interests and doctorate. Sociologica , which is the Italian Journal of Sociology, has just republished Harrison White's "Notes on the Constituents of Social Structure ". Harrison White is the Giddings Professor of Sociology at Columbia University, however his doctorate is in theoretical physics. White is very influential in network analysis and is sometimes credited with developing block-models - but I digress.

In this article White introduces the notion of a "catnet", where he brings together the ideas of categories and networks - hence catnet. The idea at the time was quite novel, yet now it seems obvious - people who are alike in some way will form networks (birds of a feather flock together). White's contribution was how to represent the various categories that might be found in a network. Further by bringing together network analysis and categories like race, gender, or education, he was able to show that categories are an artificial construct that aid understanding but of themselves are not sufficient. ...

read more ...

I recently came across this blog-post by Marc Aafjes on what he calls the Shadow Organisation. Marc says:

"By connecting various participants across the company around the execution of our knowledge strategy we're cultivating a meta network - the shadow organisation - that enables the company to enhance the value we derive from the knowledge we have. Framing knowledge management in economic terms, the shadow organisation in effect is ‘making the market for knowledge' by connecting otherwise disparate parts of the company around knowledge needs. This shadow organisation consists of the change agents that help us execute the knowledge strategy and embed sustainable change in all parts of the company".

Now what Marc is doing is by no means new - he's weaving a network to build a community of practice! What he has done is come up with a clever name that markets his network weaving initiative. ...

read more ...

I purchased this book on the recommendation of a colleague. Unusually for me it has taken many months to read, and I have found it a hard slog - I simply could not maintain my interest. The content is dense and at times challenging.

The authors bring together several theories to come up with an integrative framework to research communication networks. By combining several approaches they seek to move from descriptive and exploratory techniques to inferential and confirmatory models - this was the attraction of the book for me; unfortunately in the end I wasn't completely convinced. That said I do agree that networks should be examined on multiple levels and that a multi-theoretical approach has considerable merit.

read more ...

Sometime ago I posted some commentary about this study by Doctor Nicholas Christakis and Doctor James Fowler, who used network analysis to aid understanding of obesity. Their results were published in the New England Journal of Medicine. It was an interesting study because it looked at 12,067 subjects over a 32-year period. Some of the results were truly frightening - if you have a close friend who is obese the chance of you becoming obese increases by 57%! But I digress. This time, Doctors Christakis and Fowler have used social network analysis techniques to aid understanding of the collective dynamics of smoking.

Again it is a very interesting study. Given the population is 12,067 subjects over a 32-year period it appears to be the same group used in the obesity study. A couple of findings were really interesting.

• First, smoker-clusters extend to three degrees of separation.
• Second, the size of the smoker-clusters remained the same across time, suggesting that whole groups of people quit at the same time.
• Third, smokers are found in the periphery of the total social network, suggesting smokers are increasingly being marginalized by their peers, family and co-workers. ...

read more ...