Creator Graph

What the Creator Graph Is

The creator graph is Repute AI’s model of the Telegram creator ecosystem as a network structure. It represents the relationships between Telegram channel operators, independent creators, and the narratives they carry.

Understanding this graph is essential to understanding how NSM (Narrative Strength Model) works and why creator density is such a powerful structural signal.

Graph Structure

The creator graph is a weighted, directed graph:

Nodes = Telegram channels and creators (each channel is a node; creators who operate multiple channels are a single node with multiple channel associations)
Edges = Shared narrative connections between creators (a directed edge from Creator A to Creator B means Creator A’s content references or propagates Creator B’s narrative)
Edge weight = Strength of narrative connection (frequency × recency of shared narrative propagation)

Illustration

[Creator A] ──────── narrative X ─────────► [Creator B]
     │                                            │
     │ narrative X                    narrative X │
     ▼                                            ▼
[Creator C] ◄──────── narrative X ────────[Creator D]

In this structure, narrative X is propagated by four independent creators. The creator graph for narrative X has four nodes and multiple edges. This is a high-density graph — the narrative has structural depth.

Compare to:

[Project Channel] ──► narrative Y ──► [Paid Promoter A]
                                 └──► [Paid Promoter B]

Narrative Y is propagated by only the project itself and two paid promoters. The graph is sparse and centralized — the narrative has no independent structural support.

Dense vs. Sparse Graphs

Graph Property	Dense Graph	Sparse Graph
Creator count	6+ independent nodes	1–3 nodes
Edge distribution	Multiple independent paths	Single centralized source
Narrative independence	Narrative exists without any single node	Narrative dies if source node stops
NSM contribution	High	Low
Structural persistence	High	Low

Dense graph = strong narrative cluster. When the creator graph around a narrative is dense — many independent creators with multiple connections — the narrative is self-sustaining. No single creator’s removal collapses it.

Sparse graph = structural fragility. When the creator graph is centralized around the project’s own channel, the narrative is entirely dependent on the project’s promotional energy.

Graph density is computed using the ratio of actual edges to maximum possible edges for the observed node count. A density above 0.4 for a cluster of 5+ creators is considered structurally significant.

How Graph Density Feeds Into NSM

The NSM CreatorDensity variable is derived directly from the creator graph:

CreatorDensity = IndependentCreatorNodes / EcosystemBaseline

Where:

IndependentCreatorNodes = number of creator nodes in the narrative graph that are not the project team, not verified as paid promoters, and have independent follower bases
EcosystemBaseline = the expected number of creators per narrative in the current ecosystem context (varies by narrative category and time period)

High CreatorDensity in the creator graph directly increases NSM, which in turn increases BPM and REP.

Graph Updates

The creator graph is updated on two cadences:

Daily full refresh: All creator-to-creator connections are re-evaluated based on content propagation in the previous 24 hours
Real-time event triggers: When a high-AVI event is detected, the creator graph for the associated narrative is immediately re-evaluated to check if new creators are entering the cluster

Anomaly Detection Probability Modeling