An independent study reference written by Dr Phuc V. Nguyen. It is not official subject material — for assessment requirements always follow your subject outline and vUWS.

Networksbeginner

What social network analysis is

Q: A human resources team has complete attribute data on every employee and finds no difference between two departments on skills, tenure or workload, yet one department consistently misses deadlines. What does a network view add?

It describes who depends on whom, which can differ sharply between two groups with identical attributes. Attributes describe actors one at a time and the two departments match on all of them. Structure is a separate layer of information. One department may route every difficult decision through a single person while the other spreads it, and that difference is invisible in a staff table but obvious in a network of who consults whom.

Q: An analyst maps a 300-person organisation by drawing every email exchange, every reporting line and every project co-membership onto one diagram. The main problem is that

three different relations have been merged, so no measure computed on the result has a clear meaning. Each relation carries its own logic. Advice, authority and co-membership behave differently and reward different measures. Merging them produces ties that mean three incompatible things at once, so any density or centrality score computed on the result is uninterpretable. Map one relation per network, then compare the networks.

Social network analysis studies the pattern of relationships between actors rather than the attributes of actors taken one at a time. Ordinary analytics asks what each customer, employee or firm looks like. Network analysis asks who is connected to whom, and treats the tie as the unit of analysis. The payoff is that structure is associated with outcomes that attributes alone do not predict. Two teams with matched skills, tenure and budgets can perform very differently, and the difference often tracks the shape of their advice structure. One is connected. The other routes every hard question through a single person. Association is the honest word here, because people choose their ties for reasons that may also drive their results. The methods come from sociology and graph theory and now run on ordinary business data.

Try it yourself

What network analysis measures

A staff list describes people one at a time. This describes who is joined to whom. Switch ties on and off and watch four measures rank the same six people, because each one answers a different question about position.

Ties switched on6 of 15

6 actors, so 6 x 5 / 2 = 15 possible pairs. Top on degree: Ana.

Circle size and shade both show degree, and the number under each name is that same figure, so nothing is carried by colour alone. Ana holds the highest score.

Size and shade the actors by

Path distance

Selected actor

Ties (6 of 15 on)

Ana-Ben12Ana-Cara10Ana-Dan2Ana-Eve5*Ana-Finn5*Ben-Cara9Ben-Dan5*Ben-Eve5*Ben-Finn5*Cara-Dan5*Cara-Eve5*Cara-Finn5*Dan-Eve3Dan-Finn5*Eve-Finn8

Weights are contact counts over a seven-day period. The six ties of the running team carry the recorded counts. The other nine pairs carry a placeholder of 5, marked with an asterisk, so they can be switched on at all. No placeholder tie is on, so every weight on screen is a recorded one.

Every actor, every measure, recomputed from the ties on screen. Σd is the summed shortest-path distance to the other five actors, in steps. Closeness is 5 divided by that sum. Harmonic is the normalised harmonic closeness, a different measure that averages 1 divided by each distance and treats an unreachable actor as a contribution of zero, with n held at 6 for the whole team rather than shrunk to a component. Betweenness counts each unordered pair once and is not normalised. Clustering is 2e divided by k(k−1) and needs at least two neighbours.
Actor	Degree	k / 5	Strength	Σd (steps)	Closeness	Harmonic	Betweenness	Clustering
Ana (selected)	3	0.60	24	8	0.625	0.767	6	0.33
Ben	2	0.40	21	11	0.455	0.617	0	1.00
Cara	2	0.40	19	11	0.455	0.617	0	1.00
Dan	2	0.40	5	8	0.625	0.700	6	0.00
Eve	2	0.40	11	10	0.500	0.633	4	0.00
Finn	1	0.20	8	14	0.357	0.467	0	—

The team is one component, so every distance is finite and standard closeness has a value for everyone. Clustering is an em dash for Finn, because a coefficient needs at least two neighbours to have any pair to check.

Ties 6 of 15 Components 1 Density 0.40 Standard closeness defined

Ana holds 3 ties, which is 0.60 once divided by the 5 other actors. Degree stops at the first step, so it cannot tell you where those ties lead. The two measures disagree here. Ana leads on degree with 3, while Ana and Dan broker the most pairs at 6. That gap is the reason a network study reports more than one centrality. Strength happens to put the same name on top here, Ana on 24 contacts, because at this setting the heaviest ties and the most ties sit with the same person.

Every figure here is recomputed from the ties currently switched on. Distances come from breadth-first search on steps, or from Dijkstra on 1 divided by the weight when weighted distance is selected, and betweenness comes from the Brandes accumulation over those same shortest paths. Betweenness counts each unordered pair once and is left unnormalised. Bridges and articulation points are found by removing the tie or the actor and recounting the components, not by pattern matching.

Why it matters

Picture a staff list. It gives you names, roles and start dates, and nothing about who actually talks to whom. Now draw a line every time one person goes to another for help. The drawing shows what the list hides. Who everyone leans on. Which two halves of the office barely meet. Who would take the knowledge out the door if they resigned. Network analysis is that drawing, made countable.

Before you read on — recall

A human resources team has complete attribute data on every employee and finds no difference between two departments on skills, tenure or workload, yet one department consistently misses deadlines. What does a network view add?

Formulas

Possible ties in an undirected group of n actors

\binom{n}{2} = \frac{n(n-1)}{2}

A group of

n

actors contains that many distinct pairs, so this is the ceiling for a simple undirected relation with no self-ties. Ten people give 45 possible ties, 100 people give 4,950. A directed relation doubles the ceiling to

n(n-1)

, because a tie running one way is separate from the return tie. The count grows with the square of the group, which is why deciding who belongs inside the study matters far more here than in ordinary survey work.

Worked examples

Scenario

A retail bank wants to know why one branch resolves complex customer complaints in two days while another of the same size, with the same training and the same systems, takes nine.

Solution

An attribute comparison finds nothing, because the attributes match. A network view asks each staff member which colleagues they consult when a case is unusual. The fast branch turns out to have several people who are consulted by others, so a hard case can be routed in one hop. The slow branch has a single long-serving officer everyone consults, and every hard case queues behind that one person. Matching on training, systems and size narrows the field of explanations without settling it, because staff also choose whom to consult. The structural reading is a strong lead rather than a proven cause: spread the knowledge that currently sits in one head, then measure whether resolution time moves.

Scenario

A software firm is choosing which 20 of its 400 staff to include in a pilot of a new internal tool.

Solution

Picking the 20 keenest volunteers optimises enthusiasm and nothing else, because volunteers cluster together and talk mostly to each other. A network view maps who asks whom for technical help, then picks people spread across different parts of that structure. The pilot then reaches 20 separate pockets of the firm rather than one. This is the everyday value of the method for an analyst. It turns "who should we target" from a guess into a measurable property of the data.

Common mistakes

✗Social network analysis is about social media. The word social refers to relations between actors, not to platforms. The same methods map supply chains, board interlocks, payment flows, co-authorship and which machines on a factory floor pass work to each other.
✗Network analysis replaces attribute analysis. It complements it. Attributes tell you what an actor is, structure tells you where the actor sits, and most useful models use both. A referral model that ignores tenure will be as weak as one that ignores position.
✗The network picture is the analysis. A drawing is a first look, and past roughly 50 actors it becomes an unreadable ball of wool. The analysis is the set of measures computed on the underlying matrix, which is why the numeric definitions matter more than the layout.
✗Any set of connections is a network worth analysing. A network only means something once you have named one specific relation and one specific boundary. Mixing friendship, reporting lines and email traffic into a single set of lines produces a picture that means nothing.

Revision bullets

•Unit of analysis is the tie, not the individual actor
•Structure is associated with outcomes that attribute data alone does not capture
•Actors can be people, teams, firms, accounts, products or documents
•Roots in sociometry and graph theory, now applied to ordinary business data
•Possible undirected ties grow as n(n-1)/2, directed ties as n(n-1), so the boundary decision is critical

Quick check

An analyst maps a 300-person organisation by drawing every email exchange, every reporting line and every project co-membership onto one diagram. The main problem is that

Connected topics

What Analytics Is Analytics in Practice Actors and Ties Degree Centrality Bridges and Cliques Network Study Social Analytics Web Mining

More in Networks

Weighted Ties Closeness Betweenness Density

Sources

Freeman (2004)
Freeman, L. C. The Development of Social Network Analysis: A Study in the Sociology of Science. Empirical Press, 2004.
Traces the field from early sociograms through to modern structural analysis.
Borgatti et al. (2009)
Borgatti, S. P., Mehra, A., Brass, D. J., & Labianca, G. "Network Analysis in the Social Sciences." Science, 323(5916), 892-895, 2009.
Survey of how network analysis moved from sociology into management, physics and computing.
Wasserman & Faust (1994)
Wasserman, S., & Faust, K. Social Network Analysis: Methods and Applications. Cambridge University Press, 1994.
The standard methods reference for the definitions used throughout this cluster.

How to cite this page

Dr. Phil's Quant Lab. (2026). What social network analysis is. Business Analytics Atlas. https://phucnguyenvan.com/analytics_atlas/concept/ba-what-is-sna