Dataset for: Forecasting and Explaining Ethanol Price Through Shapley Additive Explanations: Evidence From Two Decades of Panel Agroeconomic Data in Brazil

This dataset contains multivariate time series used to analyze and forecast ethanol price dynamics in Brazil within a sugarcane-based bioenergy system. The data integrate market, macroeconomic, policy-related, and biological feedstock variables to support predictive modeling and explainability analyses. The dataset covers the period from May 2004 to November 2023, with a weekly temporal resolution. The target variable is the ethanol price, and explanatory variables include: Biological feedstock indicators: sugarcane quality (TRS – Total Recoverable Sugars), harvest period Energy market variables: gasoline prices, international petroleum prices Macroeconomic variables: inflation rate, exchange rate Policy-related variables: Brazilian VAT Trade indicators: ethanol exports Agricultural market variables: domestic and international sugar prices All variables were obtained from publicly available official sources (e.g., national agricultural and energy agencies, economic databases, and commodity market reports). The dataset includes processed and standardized time series used for machine learning modeling, along with documentation describing each variable, its unit of measurement, source, and preprocessing steps. This dataset supports research on bioenergy market dynamics, sustainability-related economic modeling, and explainable artificial intelligence applied to agricultural and energy systems.