geostep.power.run_power_analysis

geostep.power.run_power_analysis(historical_data: DataFrame, geo_col: str, date_col: str, kpi_col: str, effect_sizes: List[float] | None = None, test_weeks_list: List[int] | None = None, pre_period_weeks: int = 8, n_sims: int = 500, alpha: float = 0.05, *, random_seed: int = 42, n_jobs: int = 1, analyzer: str = 'lift', stratum_col: str | None = None, use_permutation_test: bool = False, n_permutations: int = 9999) → DataFrame

Estimate conditional rejection rates using the actual parallel analyser.

effect_sizes are proportional interventions on historical test outcomes. stratum_col specifies frozen pre-treatment strata, with at least four eligible geos per stratum. Missing/invalid replicates are counted separately. power is unavailable if any replicate fails; conditional rejection among valid replicates is retained separately with a Wilson Monte Carlo interval. Seeds are independent of worker count. No input is mutated.