incline.trending¶

incline.trending(estimates, window_length=5, aggregation='mean', weighting='uniform', half_life=None, confidence_level=0.95, ids=None)[source]¶

Rank series by recent trend strength.

Parameters:

estimates (Mapping[str, TrendEstimate] | Sequence[TrendEstimate]) – Estimates to rank, either as a mapping from identifier to estimate or as a sequence alongside ids.
window_length (int) – How many of the most recent observations to summarize.
aggregation (Aggregation) – 'mean', 'max', 'median' or 'last'. The last value preserves its pointwise uncertainty exactly; the mean uses a conservative bound because the covariance is unavailable.
weighting (Weighting) – Weighting across the window for 'mean'.
half_life (float | None) – Observations required for exponential weights to halve. Required when weighting='exponential' and unavailable otherwise.
confidence_level (float) – Confidence level for the reported interval.
ids (Sequence[str] | None) – Identifiers, when estimates is a sequence.

Returns:

One row per series with columns id, trend, trend_standard_error, ci_lower, ci_upper, significant, uncertainty_exact, rank, sorted strongest first. The schema does not change when the input is empty.

Raises:

TypeError – If ids is supplied with mapping input.
ValueError – If an argument is outside its domain or identifiers cannot be determined.

Return type:

pd.DataFrame