:::: MENU ::::

Mark Newman’s talk

Kick off of Interdisciplinary Network Seminar

positions – may mean something, may mean nothing
questions: can spatial layout ever mean “nothing”? is it enough to say “be careful”? do we already know ho people would process spatial layouts, or does that need to be studied?

acyclic – no closed loops; ex. citation networks – a paper has to have been written before you can cite it, so you can’t go backwards in time, no loops

look at the data, notice something about it, turn it into a mathematical statement to manipulate it

Amazon network – http://www.orgnet.com/divided.html

social network = network of people
thought: not sure that’s enough to make something “social”

Mark Lombardi – art + political commentary + social networks
pierogi
story on NPR

CAIDA – cooperative association for internet data analysis – http://www.caida.org/

disease – person with 100 contacts is 100×100 more likely to spread because they’re more likely to get it and to pass it on; averged squared degrees key

thoughts: maybe dissonance is about the characterization of connections; not all connections are the same; is there some way to characterize the edges to reflect those dissimilarities and differences in magnitude within the same diagram? when do what kinds of connections matter – something to ask in interviews

graph partitioning (from CS) vs. community structure (from SS and applied math) –
GP – specified number of groups; best answer whether or not it’s good question: what?
ComStructure – clustering doesn’t predetermine number or size of groups; natural lines; surprisingly few edges = interesting

Q/A period

Gavin: asked about research on understanding meaning in space
Mark: spring layout can be misleading, but maybe we can study meaning; eigenvectors of most important dimensions – axes are meaningful but tough to read by site

STIET guy: what about multimodal networks?
Mark: assortative mixing – are people more likely to make contacts with different or similar? assortative = link similar, disassortative = link different


So, what do you think ?

You must be logged in to post a comment.