Skip to main content
Analytics processes data in three layers:

​
1. Preserve the source

CargoWise sends XML to the Analytics outbound target. The platform saves the source evidence before it builds report data. The private source archive keeps the original payload. source_data keeps its identity, parsed XML tree, scalar values, field catalog, and business ownership links. Together they support audits and historical replay. A delivery sent only to the Stamping target does not populate Analytics.

​
2. Build current business data

The Analytics worker identifies supported records, fields, and relationships. It writes them to the canonical cargowise model for shipments, quotes, charges, parties, dates, and other business details. If Lambda does not recognize a scalar field, it keeps the field in source_data. The public source_fields view can expose it. If Analytics does not yet support a structured item, its evidence stays in source_data. It does not become a business row until the CargoWise and reporting contracts support it.

​
3. Publish reporting views

The reporting schema exposes the stable client contract. Most reports query this layer.
Projection is asynchronous. CargoWise acceptance means the platform saved the delivery. It does not mean the new rows are already queryable.

​
Safe historical replay

A historical replay builds a separate data generation. Reports keep reading the active generation until the replacement passes checks and becomes active. This prevents reports from reading a partly rebuilt dataset. Read Source data, CargoWise model, and Reporting for the role of each data layer.