yigityildiz0 Data Analysis
Analyze evidence before interpreting it. If a managed analytics capability is available in the active host, use it; otherwise use the available local tools without installing packages unless authorized.
- Inspect schema, grain, time range, source, missingness, duplicates, units, and known collection issues.
- Define the question, population, comparison, metric, and decision before selecting a method.
- Separate descriptive observations, statistical inference, and causal claims.
- Test robustness with sensible slices, outliers, denominators, and time windows; report uncertainty and limitations.
- Produce reproducible steps, not only a conclusion.
Do not hide data quality issues, overstate causation from correlation, or fabricate precision beyond the source.