background image
10 / Winter 2012 / RECORD
Historically there
has not been a
whole lot of data
available about
social networks
magine, for a moment, that you could
map your network of friends on Face-
book or Twitter. You would be at the
centre, linked to all your friends. In turn,
each of those friends would be linked to
their friends, some of whom would also
be your friends. The map would provide a
glimpse of your social network.
now imagine mapping all of Facebook's
800 million users in this way.
That is the task mathematics professor
Margaret-Ellen Messinger has set for her-
self. Graphing social networks is part
of what she will be researching with the
help of a $50,000 grant from the natu-
ral Sciences and Engineering Research
Council of Canada (nSERC).
Messinger says people have been studying
complex graphs for years, but the arrival of
online social networks offers an opportu-
nity to graph social networks.
"Historically there has not been a whole lot
of data available about social networks,"
she says. "With the emergence of online
social networks like Facebook, Twitter,
Flickr and so on, comes a lot of informa-
tion about the network."
Messinger says experiments in the 1960s
tracked social networks by giving one
person a letter to deliver to a stranger in
another town. It had to be passed hand-
to-hand and could only go to someone you
knew on a first-name basis. Researchers
found it took a surprisingly short number
of passes to connect two people, usually
between two and 10.
Online social networks give researchers
the opportunity for much more sophis-
ticated study of social ties and how they
evolve over time. Messinger is helping
build tools that researchers can use to help
in these studies.
She says there are currently very few pro-
posed models for online social networks
because the field is so new.
"Where I come in is looking at the
proposed models and studying the proper-
ties that they exhibit," she says. "Ideally,
in this process of learning about the pro-
posed models, I could come up with my
own model."
A well-built model could be used to make
predictions about the spread of informa-
tion and to learn more about the structure
of the network itself.
Messinger says there is the idea of domi-
nation, for example, which is finding the
minimum number of points needed to
connect to all other points on a graph.
On a social network like Twitter, that means
finding the smallest number of people needed
to connect to every member of the network.
If Lady Gaga, Justin Bieber, and Katy Perry,
who currently have the largest number of
followers, all put out the same message,
would it get to everyone on Twitter? What
if Shakira and Kim Kardashian, numbers
four and five, joined in? How far down
the list would you need to go to get the
message to everyone?
"In general, studying online social net-
works may lead to some insight into
community structure and into social influ-
ence," Messinger says.
by Aloma Jardine
ReseaRch and cReative activity
Isolated Communities
Giant Component