HierarchicalInterruptedTimeSeries.effect_summary#

HierarchicalInterruptedTimeSeries.effect_summary(*, window='post', direction='increase', alpha=0.05, cumulative=True, relative=True, min_effect=None, treated_unit=None, period=None, prefix='Post-period', **kwargs)[source]#

Return a compact summary of the population-level effect.

Reports posterior mean and HDI for mu_lift (instant), each mu_delta bin (event-study / placebo), or L/k/s (saturation).

Parameters:

  • window (Union[Literal['post'], tuple, slice]) – Not supported; passing a non-default value emits a warning and is ignored.

  • direction (Literal['increase', 'decrease', 'two-sided']) – Direction for the tail-probability column reported per row: P(effect > 0), P(effect < 0), or P(effect != 0).

  • alpha (float) – Significance level for HDI intervals; the effective HDI probability is hdi_prob = 1 - alpha.

  • cumulative (bool) – Not supported; passing a non-default value emits a warning and is ignored.

  • relative (bool) – Not supported; passing a non-default value emits a warning and is ignored.

  • min_effect (float | None) – Not supported; passing a non-None value emits a warning and is ignored.

  • treated_unit (str | None) – Not supported; passing a non-None value emits a warning and is ignored.

  • period (Optional[Literal['intervention', 'post', 'comparison']]) – Not supported; passing a non-None value emits a warning and is ignored.

  • prefix (str) – Prefix for prose generation. Defaults to "Post-period".

  • **kwargs (Any) – Reserved for forward-compatibility; not consumed by this implementation.

Returns:

Object with .table (DataFrame) and .text (str) attributes.

Return type:

EffectSummary