causalis.data_contracts.panel_did_estimate

Module Contents

Classes

PanelDIDDiagnosticData	Diagnostic payload for scalar panel DID estimators.
PanelDIDEstimate	Result contract for scalar panel difference-in-differences estimates.
CallawaySantAnnaDIDEstimate	Result contract for Callaway-Sant’Anna staggered-adoption DID estimates.

API

class causalis.data_contracts.panel_did_estimate.PanelDIDDiagnosticData(/, **data: Any)

Bases: pydantic.BaseModel

Diagnostic payload for scalar panel DID estimators.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

model_config

‘ConfigDict(…)’

unit_ids: List[Hashable]

None

d: numpy.ndarray

None

delta_y: numpy.ndarray

None

x: Optional[numpy.ndarray]

None

covariate_names: List[str]

‘Field(…)’

propensity_score: numpy.ndarray

None

control_outcome_evolution: numpy.ndarray

None

treated_weights: numpy.ndarray

None

control_weights: numpy.ndarray

None

influence_scores: numpy.ndarray

None

gamma_hat: numpy.ndarray

None

beta_hat: numpy.ndarray

None

overlap: Dict[str, Any]

‘Field(…)’

balance: pandas.DataFrame

‘Field(…)’

cluster_scores: Optional[pandas.Series]

None

class causalis.data_contracts.panel_did_estimate.PanelDIDEstimate(/, **data: Any)

Bases: pydantic.BaseModel

Result contract for scalar panel difference-in-differences estimates.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

model_config

‘ConfigDict(…)’

estimand: Literal[ATT]

‘ATT’

model: str

None

treatment_start: causalis.data_contracts.panel_data_did.TimeLike

None

pre_time: causalis.data_contracts.panel_data_did.TimeLike

None

post_time: causalis.data_contracts.panel_data_did.TimeLike

None

att: float

None

se: float

None

ci_lower: float

None

ci_upper: float

None

p_value: float

None

is_significant: bool

None

alpha: float

None

n_units: int

None

n_treated: int

None

n_control: int

None

treatment_mean_delta: float

None

control_mean_delta: float

None

outcome: str

None

treatment: str

None

covariates: List[str]

‘Field(…)’

cluster_col: Optional[str]

None

inference: Literal[influence, clustered_influence]

‘influence’

diagnostic_data: Optional[causalis.data_contracts.panel_did_estimate.PanelDIDDiagnosticData]

None

created_at: datetime.datetime

‘Field(…)’

property value: float

CausalEstimate-style alias for the ATT.

property std_error: float

Readable alias for the influence-function standard error.

summary() → pandas.DataFrame

Return a compact scalar panel-DID summary table.

class causalis.data_contracts.panel_did_estimate.CallawaySantAnnaDIDEstimate(/, **data: Any)

Bases: pydantic.BaseModel

Result contract for Callaway-Sant’Anna staggered-adoption DID estimates.

Initialization

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

model_config

‘ConfigDict(…)’

estimand: Literal[average_post_effect]

‘average_post_effect’

model: str

None

estimator: str

None

control_group: str

None

anticipation: int

None

base_period: Literal[universal, varying]

None

include_pre_periods: bool

None

alpha: float

None

att_gt: pandas.DataFrame

None

aggregates: Dict[str, pandas.DataFrame]

None

support: pandas.DataFrame

None

skipped_cells: pandas.DataFrame

None

outcome: str

None

treatment: str

None

unit_col: str

None

time_col: str

None

covariates: List[str]

‘Field(…)’

cluster_col: Optional[str]

None

inference: str

None

diagnostics: Dict[str, Any]

‘Field(…)’

created_at: datetime.datetime

‘Field(…)’

property att: float

Simple overall post-treatment effect, weighted by treated cohort-time observations.

property value: float

Alias for the simple overall post-treatment effect.

property se: float

Standard error for the simple overall post-treatment effect.

property std_error: float

Alias for the simple overall post-treatment effect standard error.

property ci_lower: float

property ci_upper: float

property p_value: float

property diagnostic_data: Optional[Dict[str, Any]]

Compatibility alias for callers moving from older DID result objects.

aggregate(kind: Literal[simple, cohort, calendar, event]) → pandas.DataFrame

Return an aggregate table: simple, cohort, calendar, or event.

event_study() → pandas.DataFrame

Return dynamic effects by event time.

summary() → pandas.DataFrame

Return a compact summary of the average post-treatment effect.