Data and Code for: Intraday Periodicity in Ethereum's Transaction Fee Market: Empirical Evidence, Institutional-Hour Origins, and Forecasting Its Persistence Through Protocol Evolution

Data and code accompanying the manuscript “Intraday Periodicity in Ethereum’s Transaction Fee Market: Empirical Evidence, Institutional-Hour Origins, and Forecasting Its Persistence Through Protocol Evolution,” submitted to Technological Forecasting and Social Change (Elsevier). This study investigates whether Ethereum’s consensus throughput exhibits a recurring intraday cycle, whether the quietest periods (“Dead Zones”) coincide with the lunch-break hours of major global financial centers (London, New York, Hong Kong), and whether this cycle is likely to persist as the underlying protocol evolves. Using a full calendar year (2025) of granular, 10-minute-resolution Ethereum transaction and gas-price data, we apply Seasonal-Trend decomposition (STL) and non-parametric hypothesis testing (Mann-Whitney U, Cohen’s d) to identify empirically-validated low-cost transaction-scheduling windows, then decompose the documented pattern into behaviorally-anchored and protocol-anchored components to forecast its persistence through Ethereum’s forthcoming supply-side protocol upgrades. Repository contents data/ — two CSV files: 10-minute-resolution transaction counts and ETH transfer volume (365 × 144 = 52,560 rows), and gas-price/block-fullness metrics averaged across the year (144 rows). code/ — the full analysis pipeline (STL decomposition, Dead Zone/Peak Hour classification, statistical testing, sensitivity analysis, figure export) and an interactive Streamlit dashboard. queries/ — the original Dune Analytics (DuneSQL) queries used to extract the raw data from the Ethereum mainnet, for full reproducibility from the canonical on-chain source. Data provenance: raw data were queried from the Ethereum mainnet via the Dune Analytics blockchain indexer (ethereum.transactions and ethereum.blocks tables), covering 2025-01-01 to 2025-12-31 inclusive. License: code (code/, queries/) is released under the MIT License; data (data/) is released under Creative Commons Attribution 4.0 International (CC BY 4.0).