Explore and Analyze Data
For: Analytics Readers and Writers
Use this chapter for read-only exploration and analysis. Your connection needs analytics read access, and you need a Reader or Writer role for the selected Analytics Project.
Common tasks
Section titled “Common tasks”- Discover available datasets, tables, and columns.
- Understand dataset descriptions and sources.
- Ask questions in natural language.
- Filter, group, compare, and summarize data.
- Review result tables and charts.
- Check assumptions, source coverage, and incomplete results.
- Export or continue working with a result.
A reliable query workflow
Section titled “A reliable query workflow”- Ask the AI agent to confirm the Analytics Project.
- Discover the available schemas, tables, and columns.
- Clarify ambiguous terms such as sample count, latest, or missing.
- Dry-run the query before execution.
- Execute it in read mode with a reasonable result limit.
- Review the result, exact query, assumptions, source tables, row count, and audit ID.
Example requests
Section titled “Example requests”What datasets are available in the Respiratory Cohort Analytics Project? Describe their tables, columns, and sources without changing anything.
Count measurements by month and measurement name for 2026. Dry-run the query first, then show the exact SQL and query audit ID with the result.
Which sample identifiers occur more than once in the measurements table? Explain the duplicate rule you used.
Treat charts and summaries as views of the returned data, not as additional evidence. If a result is truncated or a definition is uncertain, resolve that limitation before drawing a conclusion.