Social Network Analysis for Startups

[ Note: I’ve submitted this review for Social Network Analysis for Startups on O’Reilly’s site for the book also ]

The book is read very quickly if you decide not to work on the examples. Therefore it is a nice introduction to the subject, especially for people who do not want to go through Sociology or Graph Theory books.

It has very many typographical errors. This is the book from O’Reilly that I have submitted the most typos ever.

My major concern with the book is that although it uses NetworkX for an introduction to Social Network Analysis for Startups, the authors themselves say that NetworkX is not good for say 2000 nodes and above. And yet we are in an era where Startups get considered seriously after acquiring hundreds of thousands of users. And no the final chapter on Big Data does not really help out because it is not in the same pace as the previous ones. For example where would I go to find centralities for a 200K node network since NetworkX does not cut it? This is what the intended audience of the book wants in the final chapter. Or so I feel.


networks of networks

In John Gall’s “Systemantics” two laws that play important role are stated:

  • Systems tend to expand to fill the known universe, and
  • Every system is part of a larger system*

So when I read this tweet by Steven Strogatz:

Interdependent networks, aka “networks of networks” = the next big thing in network theory? …

I was tempted to rephrase them as:

  • Networks expand to fill the known universe
  • Every network is part of a larger network (remember there is no air gap, only different kinds of latency)

Gall’s laws never stop to amaze me.

[*] – This statement actually belongs to Grady Booch who uses it while discussing Gall’s Laws of Systemantics in his “On Architecture” podcasts.

Metcalfe’s Law before Metcalfe

I copy from the preface of “Structuring Complex Systems” by J. N. Warfield:

“Since the structuring of systems has largely been done in an ad hoc way in the past, it may seem that a theory designed to permit this process to become more explicit, and to be carried out with machine assistance, would be superfluous. […] when the number of elements to be considered is large, the number of interactions to be considered is at least comparable to the square of the number of elements. The logistics of dealing with so many interactions is by itself an inhibiting factor in conducting a studied structuring exercise and in manipulating the perceived relations.”

My beloved N2 network effect pattern has been noticed before then. But it still seems that Metcalfe was the first to attach value to it.

New eBooks on Graph Theory

My twitter stream and my INBOX brought to my attention two new books on Graph Theory:

  • Graph Theory and Complex Networks: An Introduction” by Maarten van Steen. It is very interesting to note that this book is also available electronically as a personalised PDF. As the author notes: “When you write a book containing mathematical symbols, thinking big and acting commercially doesn’t seem the right combination. I merely hope to see the material to be used by many students and instructors everywhere and to receive a lot of constructive feedback that will lead to improvements. Acting commercially has never been one of my strong points anyway”.
  • The other book is the fourth edition of Reinhard Diestel’s “Graph Theory“. This book is also available electronically in different formats. I bought the student edition for €12.50 (offer expires in Aug 15, 2010).

PS: On a side-note I decided to buy a BeBook Mini

La Ola

From page 25 of Connected:

A group of physicists who usually study waves on the surface of liquids were sufficiently intrigued that they decided to study a collection of filmed examples of La Ola in enormous soccer stadiums; they noticed that these waves usually rolled in a clockwise direction and consistently moved at a speed of “twenty seats per second”.*

Damn! I have participated numerous times in such waves and never, ever thought of that!

[*] – “Mexican waves in an excitable medium“, Nature 419, 131-132 (2002).

The story of a lost manuscript

In “The development of social network analysis” (for which I have blogged too) Linton C. Freeman, among other things, tracks the efforts of different scientists to lay a mathematical foundation for SNA. For two such efforts he writes:

“both Fararo (circa 1964) and I separately set out to specify the common mathmatical properties of all these seemingly different studies. Fararo circulated but never published his paper. Mine was presented several times and eventually published, but not until twenty-five years later.”

The unpublished manuscript in question was entitled “Theory of Webs and Social Systems Data“. I contacted Professor Fararo for the unpublished manuscript. He told me that he had lost his copy and that I might be lucky by asking Professor Freeman, which I did. When I contacted Professor Freeman he was away from home, but promised to look for it. Indeed about a week later he found the manuscript, had it scanned and emailed it to me. Like I told to my wife who is an archaeologist, I think this is what it feels when they (archaeologists) make a discovery.

Prior to writing this blog post, I told this story to two friends of mine. Funnily enough they asked me the same question:

– Name one Greek Professor who (a) would answer to your email and (b) would go into all that effort to locate something written circa 1965-1966 and send it to you.

This humble blog post stands to publicly thank both Professors for their kind replies and help.

Update: After getting permission, I uploaded the document on Scribd.