how connected are you?

Read a paper today (Bachnik, Szymczyk et al. 2005) which analysed blog networks and found that blog networks share characteristics with complex networks (cliques, 'small worlds' etc). Using graph theory the authors examined the linking patterns of approx 166500 blogs of three different Polish blogging services. They distinguished two different kinds of connections between blogs - weak (idols and fans) and strong (friends) . A connection was considered weak when there was only one 'edge' between blogs, going either from blog A to blog B or vice versa. A connection was considered strong when the link was reciprocated. They also measured average 'cliquity' (great term) for each service. Another great term introduced was eminence grise - while idol is a sociometric structure which describes a person who got a large number of positive choices, though making small number of choices by itself (i.e. lots of blogs linked to it, while it linked to few) - the person of eminence grise is the person that idol chooses to link to - the idols idol! I love it! So are you an idol, a person of eminence grise or unconnected like poor me..... Other basic sociometric structures presented were diads and triads.

I found the most interesting finding (apart from the sheer number of blogs - then again Poland has a popluation of around 38 million) was that 'sparsity' was the most apparent property of the examined blogs networks - most of the blogs - about 90% - were not connected with others at all. This was way higher than I would have expected. No surmising was done in the paper as to why this might have been so. I read somewhere at some stage (how's that for vague reference....) a theory put forward for why blogs might not link to other blogs - simply that some blog creation/editing tools didn't make linking to urls easy to do for non-techies. A simple explanation but a very plausible one I thought at the time. This is probably becoming less of a factor as blogging tools become more sophisticated.

Yet another article not that relevant to my research as I hadn't planned to do link analysis, but I couldn't resist the terminology......sad I know.....

Bachnik, W., S. Szymczyk, et al. (2005). "Quantitive and sociological analysis of blog networks." Acta Physica Polonica B 36(10): 3179-3191