Researchers at the City College of New York (CCNY) have found a way to identify just who is the most likely to spread a disease or generate the most-read status updates in a human network.
Both because it's hilarious and because they share the same mechanism of action, social media phenomena such as Facebook and the spread of face-melting diseases were studied using the same method by the CCNY scientists - k-shell decomposition.
To put it simply, the CCNY team looked at four networks of individuals and examined how information or infectious diseases spread throughout the groups. The groups used included members of LiveJournal.com, inpatients at a Swedish hospital, e-mail contacts in a University CompSci server and a group of adult entertainment stars. We'll let you draw your own conclusions as to which were studied for information spreading and which for infectious diseases.
While common wisdom holds that someone near the "center" of the group, either geographically or virtually, should be the source of greatest proliferation, this was not borne out to be the case, as the CCNYers quickly discovered. Instead of simply looking at "maps" of the data, the team went through a series of "cuts" to determine which links in the network were the strongest. First, all network nodes with only one link were cut out of the picture. Only have 1 LiveJounral friend or know one other adult starlet? Cut!
The nodes that were left were assigned a "K-value" of 1. The process was then repeated until all of the remaining nodes with only one link were cut, and a higher value was assigned. This continued until all one-link nodes in the entire structure had been removed, leaving only the best and brightest at the center - those with either the most friends, most inpatient care hours, or that simply had a dedication beyond mere words to entertaining the adult community.
What researchers found was that if information or infection was released at a high k-value node, it spread extremely quickly, moreso than if it were given to a central node with a lower k-value.
Bat White Noise Syndrome Infection Map: Not sure how big the node group is for that one.
While this might seem like a way to poke fun at both adult entertainers and those who don't have a lot of e-mail contacts, the implications are a little more substantial. In the case of an epidemic or pandemic, it would be worth knowing just who posed the greatest threat of infection to a city or state, so that they could be shot first.
Sorry - so they could be given a shot first.
Similarly, if there was a need to communicate information rapidly, having access to a high k-shell value node would be very worthwhile.
This type of physical and virtual application of k-shell deconstruction is highly theoretical, but raises interesting questions about the speed at which an STD could travel from Mark Zuckerberg over the Internet pipelines.