pymargins.at_levels¶

pymargins.at_levels(variable: str, *, levels, label_fmt: str | None = None, **fixed)¶

Build one scenario per level of a categorical/binary variable.

Useful for passing to predict() to get predictions at every level, or as input to custom contrast matrices.

Parameters:

variable (str) – Variable name.
levels (list or tuple) – Levels to generate scenarios for.
label_fmt (str, optional) – Format string for labels. Defaults to "{var}={val}".
**fixed – Additional variables held constant.

Returns:

scenarios

Return type:

list[dict]

Examples

>>> scenarios = at_levels("region", levels=["north", "south", "east"])
>>> m.predict(scenarios=scenarios)

pymargins