🎯 DAG-nabit: Business Strategy Simulation

DAG-nabit is an interactive business strategy simulation dashboard designed to demonstrate the power of causal inference in marketing and customer targeting. It bridges the gap between simple A/B testing and advanced uplift modeling to help businesses optimize their strategy and maximize ROI.

---

🚀 Key Features

1. Ground Truth Simulation

• Customer Segmentation: God-mode view into 4 distinct customer types: Loyalists, Persuadables, Sleeping Dogs, and Lost Causes.
• Feature Distributions: Interactive histograms showing how observable customer features (age, activity, account value) correlate with hidden personas.

2. Experimentation Framework

• Randomized A/B Test: Power analysis and conversion lift tracking for clean experiments.
• Selection Bias Demonstration: Run "Naive" tests that target only highly active users to see how selection bias inflates marketing metrics.
• The A/B Test Paradox: Visual contrast between true causal lift and biased observations.

3. Causal Inference Methods

• Propensity Score Matching (PSM): Rebalancing biased groups to mimic randomized trials.
• Double Machine Learning (DML): Leveraging ML to isolate treatment effects in complex, non-linear environments.
• Uplift Modeling: Predicting individual-level treatment effects to identify "Persuadables" and avoid "Sleeping Dogs."

---

🛠️ Tech Stack

• Frontend: Streamlit
• Analytics: Pandas, NumPy
• Visualization: Plotly
• Causal ML: Scikit-Learn (custom wrappers), logic for PSM and DML.

---

📁 Project Structure

├── src/
│   ├── generators/       # Simulation engine & data generation
│   ├── experiments/      # A/B testing & Multi-arm bandit logic
│   └── causal_analysis/  # PSM, Double ML, and Uplift models
├── notebooks/            # Deep-dive walkthroughs & theory
├── tests/                # Unit tests (In Development)
├── app.py                # Main Streamlit dashboard
└── brand_guideline.md    # Design system and UI specs

---

🏗️ Getting Started

Prerequisites

• Python 3.10 or higher
• pip

Installation

1. Clone the repository:

   git clone https://github.com/archeltaneka/DAG-nabit.git
   cd DAG-nabit

2. Install dependencies:

   pip install -r requirements.txt

3. Run the dashboard:

   streamlit run app.py

---

🧪 Testing

The unit test suite is currently under development. You can run existing checks using:

python -m pytest tests/

---

📜 License

MIT License © 2025 Archel Taneka

⚙️ Want to contribute?

PRs, suggestions, and issues are welcome.