Case Studies

• DoWhy-The Causal Story Behind Hotel Booking Cancellations
  • Data Description
  • Feature Engineering
  • Calculating Expected Counts
  • Step-1. Create a Causal Graph
  • Step-2. Identify the Causal Effect
  • Step-3. Estimate the identified estimand
  • Step-4. Refute results
  • Comparing Results with XGBoost Feature Importance
• Estimating the effect of a Member Rewards program
  • I. Formulating the causal model
  • II. Identifying the causal effect
  • III. Estimating the effect
  • IV. Refuting the estimate
• DoWhy example on ihdp (Infant Health and Development Program) dataset
  • Loading Data
  • 1.Model
  • 2.Identify
  • 3. Estimate (using different methods)
  • 4. Refute
• DoWhy example on the Lalonde dataset
  • 1. Load the data
  • Run DoWhy analysis: model, identify, estimate
  • Sanity check: compare to manual IPW estimate
• Applying refutation tests to the Lalonde and IHDP datasets
  • Import the Dependencies
  • Loading the Dataset
  • Infant Health and Development Program Dataset (IHDP)
  • Lalonde Dataset
  • Step 1: Building the model
  • IHDP
  • Lalonde
  • Step 2: Identification
  • IHDP
  • Lalonde
  • Step 3: Estimation (using propensity score weighting)
  • IHDP
  • Lalonde
  • Step 4: Refutation
  • IHDP
  • Lalonde
• Lalonde Pandas API Example
  • Getting the Data
  • The causal Namespace
  • The do Operation
  • Treatment Effect Estimation
  • Specifying Interventions