Data and code for: Beyond Deterministic Forecasting: Validating Machine-Learning and Stochastic Contingency Models for Ready-Mixed Concrete Prices in Thailand

Raw TPSO construction-material price reports (January 2017 to September 2026), macroeconomic drivers (EPPO retail diesel, Brent crude, BOT exchange and policy rates, TPSO consumer price indices) and Python code that reproduce every table and figure of the paper. The code models ready-mixed concrete list prices as a jump process, compares seven models (GBM, compound Poisson, an XGBoost revision hazard, Merton jump-diffusion, GARCH and Markov switching) in a rolling-origin backtest with proper scoring rules, and produces 20-year contingency bands in nominal and real terms.