influence-propagation

influence-propagation

Layer 7: Interperspectival Network Analysis and Influence Flow

Version: 1.1.0 (music-topos enhanced) Trit: -1 (Validator - verifies influence patterns) Bundle: network

Overview

Influence-propagation traces how ideas, topics, and behaviors spread through social networks. It extends bisimulation-game with second-order network analysis, measuring reach multipliers and idea adoption rates.

Enhanced Integration: Condensed + Sheaf NNs

Cellular Sheaf for Influence Flow

# lib/influence_propagation.rb
module InfluencePropagation
  def self.trace_idea_adoption(idea:, origin_user:, network:, seed: 0x42D)
    # Use condensed stacks for network structure
    stack = WorldBroadcast::CondensedAnima.analytic_stack(
      network.map { |n| n[:id] }
    )
    
    # Convert to cellular sheaf for diffusion analysis
    sheaf = WorldBroadcast::CondensedAnima.to_cellular_sheaf(stack)
    
    # Trace diffusion through Laplacian
    adoption_timeline = []
    sheaf[:edges].each do |edge|
      if idea_present?(edge[:src], idea)
        adoption_timeline << {
          user: edge[:tgt],
          via: edge[:src],
          confidence: sheaf_similarity(edge)
        }
      end
    end
    
    {
      adoption_timeline: adoption_timeline,
      adoption_rate: adoption_timeline.size.to_f / network.size,
      key_amplifiers: find_amplifiers(adoption_timeline)
    }
  end
  
  def self.second_order_network(center_user:, depth: 2)
    # Profinite approximation for network layers
    direct = get_direct_connections(center_user)
    second = direct.flat_map { |d| get_direct_connections(d) }.uniq
    
    {
      direct_network: direct,
      second_order: second - direct,
      reach_multiplier: second.size.to_f / [direct.size, 1].max
    }
  end
end

DuckDB Network Schema

CREATE TABLE network_nodes (
    user_id VARCHAR PRIMARY KEY,
    username VARCHAR,
    interaction_count INT,
    first_seen TIMESTAMP,
    last_seen TIMESTAMP,
    network_depth INT  -- 1 = direct, 2 = second-order
);

CREATE TABLE influence_edges (
    edge_id VARCHAR PRIMARY KEY,
    source_user VARCHAR,
    target_user VARCHAR,
    edge_type VARCHAR,  -- 'follow', 'reply', 'repost', 'quote'
    weight FLOAT,
    created_at TIMESTAMP
);

CREATE TABLE idea_adoptions (
    adoption_id VARCHAR PRIMARY KEY,
    idea_fingerprint VARCHAR,
    user_id VARCHAR,
    adopted_at TIMESTAMP,
    confidence FLOAT,
    via_user VARCHAR  -- who they learned from
);

-- Reach multiplier query
WITH direct AS (
    SELECT COUNT(DISTINCT target_user) as direct_reach
    FROM influence_edges WHERE source_user = ?
),
second_order AS (
    SELECT COUNT(DISTINCT e2.target_user) as second_reach
    FROM influence_edges e1
    JOIN influence_edges e2 ON e1.target_user = e2.source_user
    WHERE e1.source_user = ?
)
SELECT second_reach / NULLIF(direct_reach, 0) as reach_multiplier
FROM direct, second_order;