Analysis and Postprocessing

Comprehensive Analysis

Comprehensive post-processing analysis for DeepCausalMMM with inverse transformation. Includes all visualizations: coefficients, contributions, DAG, actual vs predicted, channel analysis. Automatically handles burn-in/padding removal from all outputs.

class deepcausalmmm.postprocess.comprehensive_analysis.ComprehensiveAnalyzer(model, media_cols: List[str], control_cols: List[str], output_dir: str = 'mmm_analysis_results', pipeline=None, auto_detect_burnin: bool = True, manual_burnin_weeks: int | None = None, config: Dict | None = None, inference: InferenceManager | None = None)[source]

Bases: object

Modernized comprehensive analyzer for DeepCausalMMM with config-driven visualizations.

__init__(model, media_cols: List[str], control_cols: List[str], output_dir: str = 'mmm_analysis_results', pipeline=None, auto_detect_burnin: bool = True, manual_burnin_weeks: int | None = None, config: Dict | None = None, inference: InferenceManager | None = None)[source]

Initialize the comprehensive analyzer.

Parameters:

  • model – Trained DeepCausalMMM model

  • media_cols – List of media column names

  • control_cols – List of control column names

  • output_dir – Directory to save outputs

  • pipeline – UnifiedDataPipeline instance for modern data processing

  • auto_detect_burnin – Whether to automatically detect burn-in weeks from model

  • manual_burnin_weeks – Manually specify burn-in weeks (overrides auto-detection)

  • config – Configuration dictionary (uses default if None)

  • inference – Modern InferenceManager instance

inverse_transform_target(y_scaled: ndarray) → ndarray[source]

Apply inverse transformation to target variable using modern pipeline.

Parameters:

y_scaled – Scaled target values

Returns:

Unscaled target values

inverse_transform_contributions(contributions_scaled: ndarray, y_original: ndarray) → ndarray[source]

Apply inverse transformation to contributions using modern pipeline.

Parameters:

  • contributions_scaled – Scaled contributions

  • y_original – Original scale target values

Returns:

Contributions in original scale

analyze_with_unified_pipeline(X_media: ndarray, X_control: ndarray, y_true: ndarray, create_plots: bool = True) → Dict[str, Any][source]

Perform comprehensive analysis using the unified pipeline.

Parameters:

  • X_media – Media data (full dataset)

  • X_control – Control data (full dataset)

  • y_true – True target values (full dataset)

  • create_plots – Whether to create visualization plots

Returns:

Dictionary with analysis results

analyze_comprehensive(X_media: ndarray, X_control: ndarray, y_true: ndarray, region_ids: ndarray, weeks: List[int] | None = None) → Dict[str, Any][source]

Run comprehensive analysis with all visualizations. Automatically removes burn-in/padding from all outputs.

Parameters:

  • X_media – Media variables [n_regions, n_weeks, n_channels] (may include padding)

  • X_control – Control variables [n_regions, n_weeks, n_controls] (may include padding)

  • y_true – True target values (scaled, may include padding)

  • region_ids – Region identifiers

  • weeks – Week labels (optional)

Returns:

Dictionary containing all analysis results (burn-in removed)

Statistical Analysis

Post-processing utilities for analyzing and visualizing DeepCausalMMM results.

class deepcausalmmm.postprocess.analysis.ModelAnalyzer(inference: InferenceManager | None = None, legacy_inference: ModelInference | None = None, scaler: SimpleGlobalScaler | None = None, pipeline=None, config: Dict | None = None, output_dir: str | None = None)[source]

Bases: object

Analyze and visualize DeepCausalMMM model results with modern class-based architecture.

__init__(inference: InferenceManager | None = None, legacy_inference: ModelInference | None = None, scaler: SimpleGlobalScaler | None = None, pipeline=None, config: Dict | None = None, output_dir: str | None = None)[source]

Initialize the enhanced analyzer with unified pipeline support.

Parameters:

  • inference – Modern InferenceManager instance (preferred)

  • legacy_inference – Legacy ModelInference instance (for compatibility)

  • scaler – SimpleGlobalScaler for proper inverse transformations (legacy)

  • pipeline – UnifiedDataPipeline instance (preferred)

  • config – Model configuration dictionary

  • output_dir – Directory to save outputs

analyze_with_unified_pipeline(model, X_media: ndarray, X_control: ndarray, y_true: ndarray, channel_names: List[str], control_names: List[str]) → Dict[str, Any][source]

Analyze model results using the unified pipeline.

Parameters:

  • model – Trained model

  • X_media – Media data

  • X_control – Control data

  • y_true – True target values

  • channel_names – Media channel names

  • control_names – Control variable names

Returns:

Analysis results dictionary

analyze_predictions(X_m: ndarray, X_c: ndarray, R: ndarray, y_true: ndarray | None = None, generate_plots: bool = True) → Dict[str, Any][source]

Analyze model predictions and generate visualizations.

Parameters:

  • X_m – Media variables [n_regions, n_weeks, n_channels]

  • X_c – Control variables [n_regions, n_weeks, n_controls]

  • R – Region indices [n_regions] (reserved; InferenceManager currently builds region indices internally)

  • y_true – Optional ground truth values

  • generate_plots – Whether to generate and save plots

Returns:

Dictionary containing analysis results and metrics

static plot_coefficients_over_time(coefficients: ndarray, channel_names: List[str]) → Figure[source]

Plot mean coefficients over time for each channel.

static plot_contribution_comparison(predictions: ndarray, actuals: ndarray | None, burn_in_weeks: int) → Figure[source]

Plot actual vs predicted revenue comparison.

static plot_waterfall_chart(marketing_contribs: ndarray, control_contribs: ndarray | None, baseline: ndarray | None, channel_names: List[str], control_names: List[str]) → Figure[source]

Create waterfall chart of contributions.

static plot_contribution_donut(marketing_contribs: ndarray, control_contribs: ndarray | None, baseline: ndarray | None) → Figure[source]

Create donut chart of contribution percentages.

DAG Postprocessing

Post-processing utilities for DAG visualization and analysis.

Deprecated since version 1.0.0: This module is deprecated and will be removed in v2.0.0. Please use the modern VisualizationManager class instead:

from deepcausalmmm.core.visualization import VisualizationManager

viz_manager = VisualizationManager(config) viz_manager.create_dag_network_plot(…) viz_manager.create_dag_heatmap_plot(…)

deepcausalmmm.postprocess.dag_postprocess.plot_dag_structure(adjacency_matrix: Tensor, channel_names: List[str] | None = None, threshold: float = 0.1, title: str = 'Media Channel DAG Structure') → Figure[source]

Deprecated since version 1.0.0: This function is deprecated. Use VisualizationManager.create_dag_network_plot() instead.

deepcausalmmm.postprocess.dag_postprocess.analyze_dag_structure(adjacency_matrix: Tensor, channel_names: List[str] | None = None, threshold: float = 0.1) → Dict[str, Any][source]

Deprecated since version 1.0.0: This function is deprecated. Use VisualizationManager for DAG analysis instead.