Data flow · Diagram Design

Role-scoped analytics pipeline

Typed commerce data crosses clear ownership boundaries: engineering lands it, data science makes it trustworthy, and analytics publishes it.

Role-scoped analytics data flow A Data Engineer ingests and stores commerce data, a Data Scientist transforms and analyzes it, and an Analyst publishes a dashboard. 01INGEST 02STORE 03TRANSFORM 04ANALYZE 05PUBLISH DATAENGINEER DATASCIENTIST ANALYTICSANALYST RAW TABLE ENG Capture Eventsshop events → batchNiFi ingest LS DB ENG Land Recordsevents · ordersObject storage DB DB SCI Clean & Modelraw → trusted tableTrino · notebooks DB TB SCI Curate Metricsconversion · revenueTrino SQL TB TB ANL Publish Dashboardmetrics → decisionsBI workspace TB FL STEPS 01Ingest 02Store 03Transform 04Analyze 05Publish DATA TYPE LSStream DBDataset TBTable FLDashboard left chip = input · right chip = output FLOW Standard handoff Focal handoff Published output

THE HANDOFF

Raw becomes analysis-ready

The focal handoff marks the moment stored records become a trusted table owned by data science.

Roles stay explicit

  • Data Engineer ingests and lands
  • Data Scientist transforms and models
  • Analyst publishes the dashboard

Payloads explain change

Left and right chips show the input and output type at every active pipeline step.