Analyze SQLite data locally on your Mac
Answer a question from an existing SQLite export without uploading it to a dashboard service or setting up a database server.
Define the question before the query
Write down the unit of analysis: one order, one customer, one event or one day. Specify date boundaries, status filters, currency and whether missing values should be included. “Sales this month” is ambiguous until those choices are explicit.
Build from individual rows to totals
- Open a stable database copy and inspect its tables and columns.
- Filter a few known records in Data to understand how values are stored.
- Use a JOIN when a related table supplies a name or category.
- Group and aggregate only after verifying the joined row count.
- Reconcile the total against a small set of underlying records.
The monthly report tutorial demonstrates this with exact answers: four paid orders total 14490 cents across two months. The accompanying join tutorial shows why a customer may appear on more than one row.
Make the result reproducible
Keep the SQL and source capture date with your report. State whether the result describes order dates or payment dates, gross amounts or a different metric. Copy a small query result for a summary; use the Data view’s explicit export scope when sharing source rows.
For repeated analysis, the optional Analytics Workbook provides supported imports, transformations, saved pipeline definitions and charts. Enable it through the plugin workflow and review writes. It is not an automatic connection to every spreadsheet, cloud service or business system.
Plan for query cost
Opening a large database does not require loading every row, but a grouped query or unindexed filter may still read a large part of a table. Cancel expensive work when necessary and narrow the question. A small LIMIT on the final result does not necessarily limit the work needed to calculate it.
Read-only reporting is free. If your next step is to import or materialize new data, review the editing and workspace requirements separately. The ordinary browser remains available throughout.
Updated