Product Analytics Data Flow

Events, consent, PII isolation, warehouse sync, and downstream analytics

01 / Sources 02 / Ingest 03 / Process 04 / Store 05 / Consume Web App browser SDK events Mobile iOS / Android events Edge API collector TLS Consent Gate policy filter PII guard Event Stream Kafka topic ordered PII Vault encrypted restricted Warehouse analytics tables curated Feature Store daily batch derived Dashboards product metrics SQL ML Model ranking job features clickstream user events app events device events identity + consent PII touch accepted events append-only identity map encrypted PII normalized facts non-PII daily aggregates batch metrics SQL read-only feature vectors derived restricted join approved only Legend primary data policy / PII async batch data store

Primary Data Path

  • • Events move left to right through source, ingest, process, store, and consume stages
  • • The hot path stays visually clear even with secondary batch flows
  • • Labels name data assets instead of generic API verbs

Sensitive Boundary

  • • Consent and PII paths are styled as security flows
  • • PII lands in a restricted vault, separate from the analytics warehouse
  • • Restricted joins are visible without implying default access

Derived Consumers

  • • Dashboards read curated facts from the warehouse
  • • Feature vectors are derived by batch from analytics tables
  • • Consumption paths stay distinct from collection and consent handling

Data-flow diagram • Built with Archify • Press T for theme and E for export