SNA Chapter 1 Lecture 1

Summary

Tanmay Chakraborty introduces social network analysis (SNA) as the study of relationships and interactions in online and offline settings. The course examines how networks are structured and operate, what properties they have, and how network analysis can be applied to real-world problems.

What counts as a network?

A network is an abstraction for representing a complex system. It can describe systems ranging from the internet and the World Wide Web to biological systems, organizations, and societies. In mathematics and computer science, networks are also called graphs.

A network is formally represented as (G=(V,E)), where:

• Nodes (also called vertices or entities) are the objects being represented.
• Edges (also called links, relations, or connections) describe relationships between nodes.

Networks can include different kinds of nodes and connections. For example, a Twitter network could include users and tweets as different node types, with connections representing following, retweeting, or posting. Nodes and edges can also have attributes, such as a user’s location or occupation.

The lecture distinguishes between several network features:

• Undirected networks: Connections have no direction, as with mutual friendship.
• Directed networks: Connections point one way. In a follower network, one user may follow another without being followed back.
• Weighted edges: Weights can represent the strength of a connection, such as how frequently two users interact. Depending on the application, weights may also be negative.
• Time-stamped edges: Connections record when a relationship or interaction occurred.
• Self-loops: An edge begins and ends at the same node.
• Parallel edges: Multiple edges connect the same pair of nodes, potentially representing different kinds of interaction.
• Hyperedges: A broader form of connection that the course will discuss later.

Unless specified otherwise, the instructor generally assumes that networks do not contain self-loops or parallel edges.

Complex networks and social network analysis

The course focuses on complex networks: systems that can be extremely large and whose structure is difficult to understand just by looking at them. They consist of nodes and edges and can have non-trivial structural properties. A social network is a simplified representation of social relationships, whether they occur online or offline.

Examples of networks include:

• Online services such as Facebook, Twitter, WhatsApp, and Instagram.
• Offline communities, clubs, workplaces, universities, and families.
• Computer and telecommunications networks.
• The internet and the Web.
• Biological systems involving interactions among cells, proteins, or neurons.

People commonly belong to several networks at once. For instance, a student might be part of a university department, a research group, a friendship circle, and a family network.

Social network analysis applies network concepts and graph theory to study relationships within online or offline societies. The field draws on data mining, statistics, algorithms, and computer systems, and can also inform the design of real-world systems.

Applications introduced

The instructor previews several applications that will be explored further in the course:

• Public health: Modeling how infectious diseases spread through social contacts, evaluating measures such as lockdowns, and identifying effective ways to distribute vaccines and limit transmission.
• Social media: Recommending friends or accounts, studying how information spreads, and analyzing or countering misinformation and fake news.
• E-commerce: Recommending products based on users’ browsing, clicking, and purchasing behavior—for example, suggesting accessories related to a purchased device.
• Web search: Representing web pages as nodes and hyperlinks as edges. The lecture previews PageRank, an algorithm for assessing page importance that is associated with the development of Google.
• Crime and online harms: Studying criminal and terrorist networks, recruitment, fraud, fake accounts, cyberbullying, and other forms of online harm.
• Scientific research: Modeling papers as nodes and citations as directed links to study how knowledge spreads, how fields develop, and how research communities interact.
• Collaboration: Building co-authorship networks in which researchers are connected when they collaborate on a paper or patent. These networks can reveal communities and patterns of collaboration.

Speakers and sources featured

• Tanmay Chakraborty — Course instructor and speaker.
• Chakraborty says some course content draws on courses offered by Stanford and various Indian institutes, and thanks the content creators. Another institution or course name in the subtitles is unclear.