Public portfolio simulation demonstrating experimentation and causal analysis skills

What This Demonstrates:

End-to-end experimentation and analytics engineering skills for product and payments use cases:

• Designing statistically sound A/B experiments with clear hypotheses
• Defining primary metrics, secondary metrics, and guardrails
• Applying causal reasoning to evaluate product changes
• Translating analytical results into concrete business decisions

Technical Implementation:

• Experiment Design: Session-level randomized A/B test
  • Control: Generic payment failure messaging
  • Treatment: Guided retry messaging with actionable next steps
• Data Layer: Synthetic session-level dataset simulating realistic payment behavior
  • Outcomes: Success, failure, recovery after soft decline
  • Context: Device type, region, returning user flag
• Analysis: Python-based analysis notebook
  • Metric computation and uplift analysis
  • Guardrail validation (retries, time-to-success, hard decline rate)
  • Statistical significance testing using two-proportion z-test
• Dataset: 50,000 simulated payment sessions with encoded treatment effects
• Note: This uses simulated data. In production, I'd add: pre-experiment power analysis, sequential testing for early stopping, and heterogeneous treatment effect analysis across user segments.

Skills Showcased:

✓ Experiment design & hypothesis formulation

✓ Metrics definition (primary, secondary, guardrails)

✓ Statistical analysis (two-proportion z-test)