Market Moods in Motion: Clustering Crypto Sentiment Dynamics

This report applies time-series clustering to the Nodiens Mood Index to uncover common sentiment regimes across 67 crypto assets and interpret collective market psychology. Using the K-Shape clustering algorithm on 21-day mood trajectories assets were grouped by similarity in their sentiment evolution, revealing three distinct behavioural regimes — Panic & Rebound (V), Uptrend (Choppy), and Steady Drawdown. The results show that sentiment across assets remains fragmented, with optimism, fatigue, and recovery coexisting across different parts of the market — a hallmark of transitional sentiment environments. The findings highlight the practical value of time-series clustering for transforming complex, unstructured sentiment data into actionable behavioural insights, supporting market monitoring, diversification, and early detection of mood shifts among crypto assets.