Data Analysis
- Inspect schema, types, missingness, duplicates, units, time zones, and sampling before calculation.
- Keep raw data immutable and make transformations explicit.
- Use robust statistics and state assumptions; distinguish correlation, prediction, and causation.
- Validate totals and joins against independent checks.
- Generate editable outputs and label filters, denominators, date ranges, and uncertainty.
- Save the script or query that produces each reported result.