diff_diff.OutcomeShape#

class diff_diff.OutcomeShape[source]#

Bases: object

Distributional shape of a numeric outcome column.

Populated on PanelProfile when the outcome dtype is integer or float (np.dtype(...).kind in {"i", "u", "f"}); None otherwise. Descriptive only — these fields surface what is observed in the outcome distribution. They never recommend a specific estimator family.

Methods

__init__(n_distinct_values, pct_zeros, ...)

Attributes

n_distinct_values

pct_zeros

value_min

value_max

skewness

excess_kurtosis

is_integer_valued

is_count_like

is_bounded_unit

__init__(n_distinct_values, pct_zeros, value_min, value_max, skewness, excess_kurtosis, is_integer_valued, is_count_like, is_bounded_unit)#
Parameters:

  • n_distinct_values (int)

  • pct_zeros (float)

  • value_min (float)

  • value_max (float)

  • skewness (float | None)

  • excess_kurtosis (float | None)

  • is_integer_valued (bool)

  • is_count_like (bool)

  • is_bounded_unit (bool)

Return type:

None

classmethod __new__(*args, **kwargs)#