Data Sources

Upload sales files. Every row is validated before anything is written.

Import sales data

One row per order line. CSV (UTF-8) or Excel .xlsx; dates as YYYY-MM-DD.

Drag a file here, or

Browse files

When importingReplace existing dataThe dashboard shows only this file. Existing orders, analyses and AI insights are removed; import history and reports are kept.Add to existing dataNew orders are added; orders that already exist are skipped, never overwritten.

File format and validation rules

Required columns: order_id, order_date, store_id, channel, product_sku, product_name, category, quantity, unit_price, unit_cost, status
Optional: line_number, discount_amount, customer_id, customer_segment, customer_region, store_name, store_region
channel: in_store, online, marketplace · status: completed, cancelled, returned.
Money uses plain numbers with at most 2 decimals; discount_amount is an absolute amount per line.
If any line of an order is invalid, the whole order is rejected — orders are never imported partially.
When adding, orders that already exist are never overwritten and the same file twice is refused.
When replacing, a file with no valid orders is refused and your existing data stays as it was.
Excel formulas and macros are never executed; .xlsm files are not accepted.

Import history

Every attempt is recorded, including rejected and failed imports.

Demo dataset: run python -m app.cli seed in backend/ to load the synthetic sample (Jan 2025 – Jun 2026).