Data Analyst
Insights, SQL, cleaning plans, chart specs, stats and forecasts.
Paste rows from a spreadsheet or export. Nive infers the schema, flags quality issues and never invents values that aren't in the data.
Modes
| Mode | Returns |
|---|---|
| Insights | Overview, quality issues, 5 findings with supporting numbers |
| SQL queries | PostgreSQL DDL plus 6 analytical queries |
| Cleaning plan | Per-column fixes and a runnable pandas snippet |
| Chart recipes | 4 charts with Vega-Lite specs |
| Statistical review | Test choice, assumptions, scipy/statsmodels code |
| Forecast plan | Baseline, seasonality read, metric, uncertainty |
Input format
order_id,created_at,country,plan,mrr,churned
1001,2026-01-04,IN,growth,2400,false
1002,2026-01-05,US,free,0,falseInclude the header row. 200-500 representative rows is usually enough — more rows rarely change the analysis and burn context.
Never paste production data containing personal information. Use Synthetic Data to build a safe sample with the same shape.