Decision Log

Track significant methodological, data, and analytical decisions that impact the research outcomes.

Format

• Date: YYYY-MM-DD
• Decision: [What was decided]
• Rationale: [Why it was decided, including alternatives considered]
• Impact: [Expected impact on the project]

---

• Date: 2023-09-25

• Decision: Generate exactly N=10,000 synthetic records.

• Rationale: This size is large enough to demonstrate stable logistic regression coefficients without causing long loading or rendering times for the Quarto website.

• Impact: Ensures the demonstration repository remains lightweight and responsive.

• Date: 2023-09-26

• Decision: Utilise logistic regression instead of a more complex machine learning classifier (e.g., Random Forest).

• Rationale: Logistic regression provides easily interpretable Odds Ratios, which are standard in epidemiological literature and easier to interpret in a demonstration context.

• Impact: Prioritizes explainability over raw predictive power for this educational synthetic example.