Social Dynamics | VectorScaleDB

Social Entity Model

Five purpose-built entity types capture the full structure of social systems — from individual agents to the cultural ideas that bind them.

SocialAgent

Individual actors

Each agent carries a behavioral state vector encoding personality traits, social status, group memberships, and current emotional state. State evolves through interactions and is compressed into behavioral regimes.

SocialGroup

Emergent collectives

Groups form, merge, and dissolve dynamically. Group state vectors capture cohesion, size, leadership concentration, and behavioral norms. Track how groups evolve from loose associations to tight-knit communities.

SocialInteraction

Relationship events

Every meaningful interaction between agents is recorded with type, sentiment, and outcome. Interaction patterns reveal alliance formation, conflict escalation, and social hierarchy emergence.

CulturalMeme

Ideas that spread

Track the propagation, mutation, and extinction of cultural ideas through populations. Meme vectors capture content, virality, adoption rate, and resistance. Watch ideas compete for mindshare in real time.

SocialPressure

Environmental forces

External pressures that shape social behavior: resource scarcity, external threats, information events, policy changes. Pressure vectors create the environmental context that drives agent decisions and group dynamics.

Unified Storage

Cross-type queries

All five entity types share the same temporal-vector format. Query across types to discover how agent behavior correlates with group dynamics, how cultural memes respond to social pressure, and how interactions cascade through networks.

Dunbar Layer Analysis

Analyze social networks through the lens of Dunbar's number — the cognitive limits on social group size that structure all human social organization.

Layered Relationships

Natural social circles

Social bonds naturally organize into concentric layers of decreasing intimacy:

Support clique (5) — closest confidants
Sympathy group (15) — close friends
Affinity group (50) — good friends
Active network (150) — meaningful contacts
Extended network (500+) — acquaintances

Query Endpoint

Automated layer classification

VectorScaleDB automatically classifies each agent's social bonds into Dunbar layers based on interaction frequency, recency, and emotional intensity. Track how individuals' social circles evolve over time, detect when key relationships shift layers, and identify agents approaching cognitive social limits.

POST /v1/social/dunbar-layers
Per-agent layer breakdown with bond strength scores
Temporal evolution of layer membership

Cascade Prediction

Predict how events propagate through social networks. When a key agent changes behavior, which groups are affected and how quickly?

Network Cascades

Ripple effect modeling

Predict social cascades before they happen. When a leader defects, a rumor starts, or a resource becomes scarce, VectorScaleDB estimates which agents will be affected, in what order, and with what intensity. Cascade predictions are based on historical interaction patterns and bond strengths.

POST /v1/social/cascade-prediction
Multi-hop propagation estimation
Confidence scoring per affected agent

Influence Mapping

Who shapes whom?

Map social influence flows across the network. Identify opinion leaders, bridge nodes connecting disparate groups, and isolated clusters resistant to outside influence. Influence maps update in real time as interaction patterns change.

POST /v1/social/influence-map
Directional influence scoring
Bridge node identification

Group Dynamics

Track group formation, cohesion, fragmentation, and dissolution as emergent properties of agent interactions.

Fragmentation Risk

Predict group splits before they happen

Monitor internal behavioral divergence within groups. When sub-factions emerge or members' behavioral vectors drift apart, fragmentation risk rises. VectorScaleDB quantifies this risk continuously, giving early warning of group instability.

POST /v1/social/fragmentation-risk
Sub-cluster detection within groups
Risk scoring with temporal trends

Cooperation Equilibrium

Stable cooperation detection

Detect when groups reach stable cooperative equilibria — states where cooperation is self-sustaining and resistant to defection. Track the conditions under which cooperation emerges, persists, or collapses across different social structures.

POST /v1/social/cooperation-equilibrium
Stability scoring for cooperative regimes
Defection vulnerability analysis

Cultural Evolution

Track how ideas, norms, and practices spread through populations, mutate, compete, and go extinct.

Meme Dynamics

Ideas as temporal vectors

Cultural memes are represented as temporal vectors that evolve as they propagate through social networks. Track adoption curves, mutation events, and competition between competing ideas. Detect when a meme reaches critical mass or when counter-memes emerge to challenge established norms. The same drift-based compression that handles sensor data captures cultural evolution — stable cultural periods compress to single segments, while periods of rapid cultural change are preserved at full resolution.

# Track cultural evolution across a population
curl -X POST https://api.vectorscaledb.com/v1/social/cultural-evolution \
  -H "Content-Type: application/json" \
  -d '{
    "population_entity_type": "SocialAgent",
    "time_range": {
      "start": "2026-01-01T00:00:00Z",
      "end": "2026-03-13T00:00:00Z"
    },
    "min_adoption_rate": 0.05
  }'

# Response includes:
#   - Active memes with adoption percentages
#   - Meme mutation events and lineage
#   - Competition scores between rival memes
#   - Predicted trajectories (growth, plateau, decline)

Use Cases

Social dynamics capabilities serve researchers, game developers, and organizations modeling complex human systems.

Gaming

Living game worlds

Populate game worlds with socially aware NPCs that form alliances, spread rumors, and react to player actions as a social network. Combine with NPC AI brain tiers for full cognitive-social simulation.

Research

Computational social science

Run agent-based social simulations at scale. Test theories of cooperation, cultural transmission, and social network evolution with millions of agents and compressed historical trajectories.

Organizations

Organizational dynamics

Model information flow, influence networks, and team cohesion within organizations. Detect emerging silos, predict knowledge transfer bottlenecks, and measure the impact of structural changes.