Analyzing Market Sentiment Through Blockchain Transaction Patterns : How can blockchain transaction patterns be analysed to understand market sentiment

Abstract

Cryptocurrency markets are volatile and lightly regulated; with conditions like that, the behavioural biases and sentiment-based trading could increase. Existing sentiment measures rely heavily on indirect textual proxies such as social media and news, which often can be unreliable and subject to manipulation. This dissertation develops a direct, data-driven framework for analysing investor sentiment from Ethereum wallet-level activity. Using more than 190 million transactions between 2020 and 2024, a structured panel of 9,750 active wallets is constructed. Transaction histories are transformed into wallet-level behavioural features capturing turnover, inflow and outflow changes, trading intensity, counterparty diversity, and activity persistence. These features are used to segment wallets into distinct behavioural archetypes through Gaussian Mixture clustering, allowing systematic classification of investors ranging from heavy-outflow distributors and accumulators to high-frequency traders and whales. Behavioural biases are then analysed through three empirical lenses. Herding is measured using transaction-based frameworks that decompose correlated trading into persistence and imitation effects. Overconfidence is proxied by portfolio turnover ratios across bull and bear regimes. Panic selling is examined through event studies of wallet reactions to large drawdowns, distinguishing immediate from delayed responses. To contextualise these behaviours, the study implements a regime classification of the Ethereum market using moving averages for comparison of wallet behaviour across bull, bear, and neutral conditions. By combining clustering with behavioural finance measures, this research builds a replicable methodology for decomposing market sentiment into heterogeneous wallet-level behaviours. The approach contributes to the academic understanding of behavioural finance in blockchain markets and provides practical implications for fintech applications, trading strategy development, and risk management.