When I started watching the TED talk below, I felt that it was wholey unoriginal (we are just getting through the period of network theory where every presentation presents a power law). However, when I got to the end I felt that there was a real insight, that the narrative was a great example of the value of data visualization and mining and that the speaker - Sean Gourley - had structured the talk (7 mins) perfectly. (via Tim)
Hey Matthew,
I have watched this thing several times since you posted it and at first my thought was it was pretty cool even if it is just a power law. But, after seeing it a couple of times it makes me really wonder about some of the conclusions too - namely how much meaning can you really derive from the slope. Just an example - Gourley says that with a higher alpha comes "Fragmented, fluid, more groups, weaker" but if you look the conflict with the highest alpha you have Northern Ireland and if you look at the time series of alpha towards the end of the presentation you MAY have some fairly weird results for the surge.
Also came across these two items you might like that where posted by someone else:
Variation of the Frequency of Fatal Quarrels With Magnitude. Lewis F. Richardson.
http://www.uvm.edu/~pdodds/research/papers/others/1949/richardson1949a.pdf
On the Frequency of Severe Terrorist Events
http://arxiv.org/abs/physics/0606007
Posted by: Ericness | January 07, 2010 at 02:30 PM
He says you need to look at group size to solve the conflict. But his data disproves that in that it bounced around. Seems his analysis of the stats isn't very valuable imho.
Posted by: Seoroi | January 10, 2010 at 12:38 PM