On-Chain Sentiment & Social Signal Correlation

ROLE: You are a crypto quantitative analyst who combines on-chain behavioral data with social sentiment indicators to build predictive models. You understand that neither on-chain data nor social signals alone tell the full story, but their correlation reveals actionable patterns.

CONTEXT: I want to build a sentiment analysis system that goes beyond simple social media mention counts. By correlating on-chain activity (transaction counts, active addresses, gas usage) with social signals (Twitter mentions, Reddit activity, Telegram group growth), I aim to identify divergences that precede price movements and detect when sentiment shifts before they manifest in price.

TASK:
1. On-Chain Activity Sentiment Indicators — Define the key on-chain metrics that reflect market sentiment. Cover active address count trends (rising = growing interest), transaction count and average value shifts, new wallet creation rate as an adoption proxy, gas price patterns on Ethereum (high gas = high demand for block space), and DeFi TVL changes as a confidence indicator. Explain the baseline ranges and what constitutes a significant deviation for each metric.
2. Social Signal Data Collection — Detail how to gather and process social sentiment data for crypto. Cover Twitter/X API for mention volume, sentiment classification, and influencer tracking, Reddit comment analysis for specific token subreddits, Telegram and Discord group growth rates and message velocity, LunarCrush and Santiment as aggregated social platforms, and Google Trends data for retail interest measurement. Explain noise filtering techniques for bot activity and spam.
3. Divergence Detection Framework — Explain how to identify meaningful divergences between on-chain and social data. Cover bullish divergence (on-chain accumulation while social sentiment is negative), bearish divergence (social hype while on-chain shows distribution), volume-sentiment mismatches (high social buzz but no on-chain activity = potential manipulation), and how to score divergence strength and assign confidence levels to signals.
4. Fear & Greed Index Construction — Walk through building a custom crypto fear and greed index using on-chain and social data. Cover the component weights: exchange net flows (20%), funding rates (15%), social volume (15%), active addresses (15%), whale activity (15%), market volatility (10%), and dominance trends (10%). Explain how to normalize each component, calculate the composite score, and backtest against historical price action.
5. Narrative Tracking & Rotation Detection — Describe how to detect narrative shifts in crypto markets using combined data. Cover identifying emerging themes from social data clustering, tracking capital flows between sectors (DeFi to AI to memes) using on-chain data, measuring narrative momentum through social mention acceleration and corresponding TVL changes, and building an early warning system for narrative rotation.
6. Alert System & Signal Integration — Design a practical alert system that combines on-chain and social signals. Specify alert types: extreme fear/greed threshold breaches, divergence signals above confidence threshold, unusual whale activity coinciding with social volume spikes, narrative rotation signals, and new token social emergence with on-chain validation. Include recommended response frameworks for each alert type.