Data and Code for "How Predictable are Augmented Reality Games? "

Weekly Google Trends search-interest data (2016–2025), analysis code, and full evaluation outputs for the manuscript "How predictable are augmented reality games? A rolling-origin forecasting benchmark of Pokémon GO and Pikmin Bloomacross four markets". The package contains: (i) raw multi-segment Google Trends retrievals and the stitched continuous weekly series for eight game–region combinations (Pokémon GO and Pikmin Bloom × Worldwide, Japan, Taiwan, Hong Kong), with full stitching documentation; (ii) a Colab notebook and an equivalent script that reproduce the rolling-origin forecasting benchmark (naïve seasonal, ARIMA(2,1,2), STL+ARIMA, ETS, Prophet; 12-week horizons; 8 folds for Pokémon GO and 5 for Pikmin Bloom), Diebold–Mariano tests, dual Ljung–Box residual diagnostics, and an ARIMA order-selection robustness check; and (iii) all resulting tables and publication-quality figures. Data and results: CC BY 4.0; code: MIT.