Some results using the information provided by you
1. Facebook Friendship Networks
By studying the friendship networks provided by you, I have detected
Many people have structured community networks which contain a number of communities of friends, each for one group of activity. One can have a group of friend from school, college, work, or form a hobby, but in most cases, people have a high school and a college community. These have a high number of intra-community connections (links), but remain somewhat isolated from each other. Some common friends create bridges between the friend groups. two distinct types of graphs.
The other extremity are people whose friends know each other. This means that all the communities overlap each other, with no distinguishable community structure.
There are also people which place themselves in-between the two categories. Some communities overlap each other, while other remain more isolated.
2. Community Generation
I have uploaded a short movie showing my community generation algorithm implemented in Gephi using animations. The algorithm, still in its infancy, is meant at reproducing the type of communities detected in the real-life Facebook friendship graphs.
In this scenario 3 communities are created, one at a time. You will see that each community forms a hollow ring, then rapidly contracts as more links are added. The same process happens for all three communities which are rejected by the layout algorithm. Finally, some additional links are added between the 3 communities so that they attract themselves as galaxies with gravity do.
3. Opinion Diffusion
This second video uses the same type of graph as described above and adds unique information to nodes (individuals).
Each individual gets and initial random opinion between 0 and 1, where 0 = red = negative opinion and 1 = green = positive opinion. Every other opinion between 0 and 1 is represented by a proportional mixture of red and green.
After the graph is created we let the opinions diffuse, which means that each individual will change his opinion based on the dominant opinion of his (1st degree) friends. Because the network is community-based, each community will come to adopt one single strong opinion (either red or green). In this case, we begin with a 53% vs 47% ratio between green and red and the society evolves towards a ~2:1 ratio. Multiple runs on different graphs will have different results as the topology influences the clustering of the individuals, and thus the success of one party or another.
Understanding Conflict of Interest Networks - Harvard - Research in Action
Article on how SNA can be used as a great tool to observe and analyze conflicts of interest and assess the risks that arise in the evolving relationships between individuals or institutions. By using this approach, one could not only analyze patterns but also understand observed behaviors in networks of individuals.