Introduction to complex systems and social network analysis

Introduction to Complex Systems and Social Network Analysis

Arif KhanComplex Systems Research Centre

The University of Sydney

Lecturer (on leave)Department of CSE, BRAC University

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

What is Social Networks

• Relationship between Social Entities

A B

CD

Nodes are Entities/ActorsCan be people, organization, countries

Edges are relationshipCan be Friendship, Transaction, Trust, Relationship

Some Social Network

• Migratory Bird Network

Some Social Networks• Relationship network in a U.S. high school. Girls are pink, boys

are blue

Some Social Network

• Ingredients Network

Why Social Network Analysis?

• Because our choice/behavior has social context, we are influenced by others

Some Research Facts

• Relationships are mostly mutual. If you like me, I will like you.

A B

• Friend of my friend will eventually be my Friend

A B

C

Some Research Facts

• Making and Maintaining friendship is costly.• Average person can maintain only 150

intimate relationship.• So this friendship network is hardly possible.

Some Research Facts

• This is a small world. On average each person of world has a friendship distance of 6

• There are on average 3.74 people in between any two facebook users.

Complex System

• Social Networks are one type of Complex System, they have all the properties discussed above.

• Complex Systems– Have large number of actors/nodes– Changes over time– They are resilient.

Centrality: finding important Nodes

• Degree Centrality

A B

C

D

E

I am So popular!!

A B

C

D

E

I am social,

I love

everyone!!

B has high in-Degree

B has high out-Degree

Centrality: finding important Nodes

• Betweenness Centrality‘A’ has high Betweenness centralityIt can control Information, can control business,Can manipulate information

Centrality: finding important Nodes

• Closeness centralityI may not have

lots of friend. But I have powerful

friend, enough for me

High (not highest) closeness centrality

Example of SNA software (Gephi)

• Analyze my Facebook network• <OPEN GEPHI outside>

Social Network Analysis as CSE perspective

• Developing efficient algorithms (graph algorithm, approximation algorithm)

• Developing efficient graph/mathematical models of real world Social Networks

• Developing data mining software, centrality measures

• And many more…

If you are interested…

• This is relative new research area.• Almost all universities have research centers

related to Social Network analysis, complex system, data mining etc.

• Complete the online course (FREE!) from www.coursera.org/course/sna

THANK YOU

Questions?

Important Links: www.coursera.org/course/snaMy Mail: [email protected]

Pictures taken from various sources, may be copyrighted. Used for Educational purpose only