Quick Start

Quick Start

# Quick Start

This guide will help you get started with **CausalPipe** by showing how to set up a basic causal analysis pipeline.

## 1. Define Configuration

First, define the configuration for your causal discovery pipeline using the `CausalPipeConfig` dataclass. You’ll specify variable types, preprocessing parameters, and methods for skeleton identification and edge orientation.

```python
from causal_pipe.pipe_config import (
    DataPreprocessingParams,
    CausalPipeConfig,
    VariableTypes,
    FASSkeletonMethod,
    FCIOrientationMethod,
    PearsonCausalEffectMethod,
)

from causal_pipe import CausalPipe

preprocessor_params = DataPreprocessingParams(
        cat_to_codes=False,
        standardize=True,
        filter_method="mi",
        filter_threshold=0.1,
        handling_missing="impute",
        imputation_method="mice",
        use_r_mice=True,
        full_obs_cols=None,
)

# Define variable types
variable_types = VariableTypes(
    continuous=["age", "income"],
    ordinal=["education_level"],
    nominal=["gender", "diagnosis_1", "diagnosis_2"],
)

# Initialize the configuration
config = CausalPipeConfig(
    variable_types=variable_types,
    preprocessing_params=preprocessor_params,
    skeleton_method=FASSkeletonMethod(),
    orientation_method=FCIOrientationMethod(),
    causal_effect_methods=[PearsonCausalEffectMethod()],
    study_name="causal_analysis",
    output_path="./output",
    show_plots=True,
    verbose=True,
)

CausalPipe also ships with a range of specialized classes for each stage of the pipeline. You can mix and match these to suit your analysis:

• Skeleton methods: FASSkeletonMethod, BCSLSkeletonMethod
• Orientation methods: FCIOrientationMethod, HillClimbingOrientationMethod
• Causal effect methods: PearsonCausalEffectMethod, SpearmanCausalEffectMethod, MICausalEffectMethod, KCICausalEffectMethod, SEMCausalEffectMethod, SEMClimbingCausalEffectMethod, PYSRCausalEffectMethod, PYSRCausalEffectMethodHillClimbing

2. Initialize CausalPipe

Initialize the CausalPipe toolkit by passing in the configuration:

from causal_pipe import CausalPipe

# Initialize CausalPipe
causal_pipe = CausalPipe(config)

3. Run the Causal Discovery Pipeline

Now, you can run the pipeline on your data:

import pandas as pd

# Load your dataset
data = pd.read_csv("your_data.csv")

# Run the causal discovery pipeline
causal_pipe.run_pipeline(data)

4. Estimating Nonlinear Mechanisms with PySR

CausalPipe can learn symbolic structural equations using PySR. Enable it by adding the pysr causal effect method:

from causal_pipe.pipe_config import PYSRCausalEffectMethod

config.causal_effect_methods = [
    PYSRCausalEffectMethod(hc_orient_undirected_edges=True)
]

causal_pipe = CausalPipe(config)
results = causal_pipe.run_pipeline(data)
print(results["pysr"]["structural_equations"])

---

For more details on the API, see the API Reference.