shreyajaiswal1604 / econml_customer_segmentation_case_study Goto Github PK

Welcome to the Dynamic Business Insights project! This repository explores the application of fundamental concepts such as causality, causation, and causal inference in the realm of statistical analysis, with a specific focus on a real-world case study: Customer Segmentation at an Online Media Company.

• Causality:

  • Understanding the relationship between cause and effect.
  • Fundamental for making predictions, guiding interventions, and optimizing strategies.

• Causation:

  • Direct influence of one variable on another.
  • Classic "counterfactual" definition.

• Causal Inference:

  • Pivotal in understanding cause-and-effect relationships.
  • Goes beyond correlation, drawing conclusions based on observational or experimental data.

Explore the strategic marketing approach of dividing a broad target market into smaller, manageable customer groups:

• Purpose:

  • Create personalized and targeted marketing strategies.
  • Design approaches resonating better with specific customer groups.

• Types of Segmentation:

  • Demographic, Geographic, Psychographic, Behavioral.

• Benefits:

  • Improved Targeting.
  • Increased Customer Satisfaction.
  • Resource Optimization.

• Examples:

  • Online clothing retailer campaigns.
  • Food delivery service promotions.

• Implementation:

  • Utilizing data analysis tools, surveys, and machine learning for pattern identification.

Explore the application of causal inference techniques and data-driven strategies in a dynamic online media company scenario:

• Objective:

  • Estimate individualized responses for personalized pricing strategies.

• Approach:

  • Causal Inference: DoWhy library for constructing and validating causal models.
  • Estimation: EconML's DML-based estimators leveraging user features.
  • Interpretation: SingleTreeCateInterpreter for summaries on responsiveness to discounts.

To dive into this project and explore the dynamic business insights, follow these steps:

git clone [email protected]:ShreyaJaiswal1604/EconML_Customer_Segmentation_Case_Study.git

cd EconML_Customer_Segmentation_Case_Study

conda create --name econml_env python=3.8
conda activate econml_env
pip install -r requirements.txt

1_Data_Preparation.ipynb: Preprocesses the raw data. 2_Causal_Inference_Modeling.ipynb: Constructs and validates causal models using DoWhy. 3_EconML_Modeling.ipynb: Estimates individualized responses using EconML's DML-based estimators. 4_Results_Interpretation.ipynb: Summarizes responsiveness to discounts using SingleTreeCateInterpreter. Execute the notebooks sequentially for a comprehensive understanding of the project.

This project is licensed under the MIT License - see the LICENSE file for details.