Community Clusters
Accounts that follow each other more than they follow the rest of the network are grouped into clusters - each assigned a distinct color. These clusters often correspond to professional circles, friend groups, or shared interests.
How 
Works
A guide to influence network analysis - how networks are built, how communities are detected, and what each exploration mode reveals.
Foundations
What are influence networks?
A social network can be represented as a graph - a structure made of nodes (accounts) and edges (follow relationships). When you map the follow connections of public figures, patterns emerge that are invisible from follower counts alone.
Accounts cluster into communities: groups of people who follow each other more densely than they follow outsiders. A footballer's network might reveal clusters of teammates, national team peers, athletes from other sports, and entertainment figures. These clusters tell you who actually exists in someone's social orbit - not just who follows them.
Some accounts sit between clusters. These bridge accounts - managers, collaborators, media figures - connect communities that would otherwise be separate. They are the conduits through which influence and information flow across different professional circles.
GraphAura maps these structures for any public X (Twitter) account, using only publicly available follow data. No login is required, and no private information is ever accessed.
Exploration Modes
Three scales of network analysis
Network Graph
Start from any public X account and explore their social world as an interactive graph. The network is built by tracing follow connections outward from the chosen account, layer by layer. Direct connections appear first; then connections-of-connections fill in the broader picture. The result is a map of the account's real social proximity - not just who follows them, but how everyone in their orbit relates to each other.
What you'll see
Mutual Connections
When two accounts follow each other (a mutual follow), that relationship is treated as significantly stronger in the layout. Mutuals appear closer together and are visually distinct, revealing the seed account's closest peers.
Network Hubs
Some accounts have an unusually high number of connections within the graph. These hubs often bridge multiple communities - they might be managers, collaborators, or public figures who connect different social circles.
AI Aura Analysis
Each account is classified by profession (such as "F1 driver", "rapper", or "CEO") and assigned a 0–100 Aura Score that captures their influence within the network. A plain-English summary describes the account's social landscape.
Side-by-Side Compare
Place two networks next to each other to see which accounts they share, where their communities diverge, and which connections bridge the two social worlds.
Under the Hood
How the graph is built
Every visualization follows the same core process - from raw follow data to an interactive, explorable graph.
01
Collecting follow data
GraphAura works from pre-populated public follow data. Follower and following lists of public X accounts are collected in advance - no live API calls happen when you view a graph. Only publicly visible profiles are included.
02
Tracing connections outward
Starting from the seed account, the network traces follow connections layer by layer: first direct follows, then their follows, and so on. At each step, the most-connected accounts are prioritized and a configurable cap keeps the graph focused and readable.
03
Weighting relationships
Each follow becomes an edge in the graph. Mutual follows - where both accounts follow each other - are given significantly more weight than one-way follows. This makes reciprocal relationships (real peers) more prominent than passive follows.
04
Detecting communities
A community detection algorithm partitions the graph into clusters: groups of accounts more densely connected internally than externally. Each cluster receives a distinct color. The process is fully automatic - no manual labeling.
05
Positioning nodes
A physics-based layout simulation positions every node. Connected accounts attract each other; all nodes gently repel. After the simulation converges, the spatial arrangement reflects real social distance - tightly-knit communities land close together, loosely connected ones drift apart.
06
AI enrichment
An AI model analyzes the network structure, classifies each account by profession and country, and generates an Aura Score (0–100) with a plain-English summary describing the account's influence landscape.
Classification
How accounts are categorized
Every account in GraphAura is classified by an AI model into a specific profession - not a generic label. An account is tagged as “F1 driver”, not just “athlete”; as “rapper”, not just “musician”; as “CEO”, not just “business”.
These specific categories are then grouped into 16 broader super-categories: athlete, musician, entertainer, influencer, media, politics, business, tech, fashion, religion, royalty, brand, organization, sports league, creator, and other. Super-categories power the Ecosystem Explorer and Panorama Mode, letting you compare entire professional worlds at once.
For multi-talented figures, the classification picks what the person is most famous for. Dwayne Johnson is categorized as “actor” (not wrestler); Kylie Jenner as “reality star” (not model). This keeps each account in one clear category, avoiding ambiguity in the graph.
Accounts are also tagged by country (ISO code) and whether they represent a real individual versus a brand or organization. The “real people only” filter in Panorama and Ecosystem modes uses this classification to focus the network on authentic individuals.
FAQ
Frequently asked questions
What data does GraphAura use?
Only publicly available follow relationships from X (Twitter). GraphAura does not access private accounts, DMs, or any non-public information. No login or account connection is ever required.
How does community detection work?
GraphAura uses the Louvain algorithm to partition the network. It identifies groups of accounts that are more densely connected to each other than to the rest of the graph. Each community is automatically assigned a distinct color.
What is an Aura Score?
A 0–100 AI-generated measure of a public figure's influence within their network, based on their position in the graph, cluster structure, and connections. It comes with a plain-English summary describing the account's social landscape.
How are categories assigned?
An AI model classifies each account by their specific profession (e.g. "F1 driver", "rapper", "CEO") based on their public profile. These specific categories roll up into 16 broader super-categories like athlete, musician, entertainer, or tech.
How fresh is the data?
Graph data is pre-populated and periodically refreshed. The most popular and recently requested accounts are updated most frequently.
What is the difference between the three modes?
Network Graph explores one account's connections outward. Ecosystem Explorer compares how entire professional categories follow each other. Panorama Mode shows the broadest view - all categories and their top accounts in one graph.
Start exploring
Head back to GraphAura to try Network Graph, Ecosystem Explorer, or Panorama Mode.