🔍 Audience Overlap Analysis

Find shared followers between accounts, rank audience similarity, and discover niche communities. Competes with SparkToro and Audiense.

Overview

Audience Overlap scrapes and compares follower lists between X/Twitter accounts to answer:

• Who follows both accounts? — shared followers, unique-to-A, unique-to-B
• How similar are their audiences? — Jaccard similarity coefficient
• Which account has the most similar audience? — rank candidates by overlap
• What are the cross-pollination opportunities? — actionable insights

Available via: CLI, MCP tools, API, and Node.js library.

Quick Start

CLI

# Analyze overlap between two accounts
unfollowx audience elonmusk jack

# Limit to 3000 followers per account
unfollowx audience elonmusk jack --max 3000

MCP (AI Agents)

Tool: x_audience_overlap
Args: { "username1": "elonmusk", "username2": "jack" }

Node.js

import {
  analyzeOverlap,
  multiOverlap,
  findSimilarAudience,
  getAudienceInsights
} from 'xactions/src/analytics/audienceOverlap.js';

// Two-account overlap
const result = await analyzeOverlap('elonmusk', 'jack', { limit: 5000 });
console.log(result.insights);
// { overlapCount: 1200, jaccardSimilarity: 0.15, ... }

// Multi-account overlap matrix (3+ accounts)
const matrix = await multiOverlap(['elonmusk', 'jack', 'naval'], { limit: 3000 });
// { matrix: [[1, 0.15, 0.08], ...], coreAudience: [...], nicheAudience: [...] }

// Find which account's audience is most similar to yours
const ranking = await findSimilarAudience('myaccount', ['elonmusk', 'jack', 'naval']);
// { rankings: [{ username: 'naval', similarity: 0.22 }, ...] }

API

# Audience overlap
GET /api/analytics/overlap?username1=elonmusk&username2=jack

Architecture

src/analytics/
├── audienceOverlap.js → Core overlap analysis + caching
└── index.js           → Re-exports

How It Works

1. Scrapes follower lists for both accounts (with in-memory caching)
2. Uses Set intersection/difference for O(n) overlap computation
3. Calculates Jaccard similarity: |A ∩ B| / |A ∪ B|
4. Generates natural-language insights

Function Reference

analyzeOverlap(username1, username2, options)

Two-account follower overlap analysis.

Returns:

{
  accountA: { username, followers },
  accountB: { username, followers },
  shared: ['user1', 'user2', ...],
  uniqueToA: ['user3', ...],
  uniqueToB: ['user4', ...],
  insights: {
    overlapCount: 1200,
    overlapPercentA: 24.0,
    overlapPercentB: 15.0,
    jaccardSimilarity: 0.15,
    analysis: 'Moderate overlap...'
  }
}

multiOverlap(usernames, options)

Pairwise overlap matrix for 3+ accounts.

Returns: { usernames, matrix, coreAudience, nicheAudience }

• matrix — 2D array of Jaccard similarities
• coreAudience — users who follow ALL accounts
• nicheAudience — users who follow only ONE account

findSimilarAudience(username, candidateUsernames, options)

Ranks candidates by audience similarity to a target account.

Returns: { target, targetFollowers, rankings: [{ username, similarity, overlap }] }

getAudienceInsights(overlapResult)

Generates natural-language insights from an overlap result object.

Returns: Array of insight strings.

clearCache()

Clears the in-memory follower cache. Use between analyses to force fresh data.

MCP Tools