Skip to main content

Usage Analytics Overview

Usage Analytics reveals how your data is actually used: which products are searched, viewed, queried, requested, subscribed to, ignored, or carrying risk signals. It combines product interaction events, access requests, subscriptions, BI/query usage, incidents, and stale grants so stewards can tell the difference between "what data exists" and "what data matters."

Why Usage Analytics?​

Without usage data, governance decisions are made in the dark:

  • Stewards can't prioritize which tables to document (most-queried vs. unused)
  • Analysts can't see popular queries for inspiration on unfamiliar tables
  • Lineage stops at the warehouse — dashboard-to-table connections are missing
  • Stale products can't be identified by actual access patterns

Usage Analytics solves these problems by bringing query-level insights directly into the catalog.

How It Works​

Usage data flows into Qarion through several complementary channels:

Product Adoption Events​

Qarion records durable product-level events when users:

  • Search for products in the catalog
  • Open product detail pages
  • Query products from the query editor
  • Request access
  • Subscribe to product updates
  • Mark product usage as ignored or stale

These events power the adoption score, product ranking, dashboard widget, and Executive Proof rollup.

BI Tool Scraping​

BI connectors scrape saved queries, chart definitions, and dashboard structures from your BI platforms. This produces:

  • Dashboard-to-table lineage — Dashboard and report nodes appear in the lineage graph
  • Saved query references — Which tables and columns are used in saved analyses
  • Chart SQL extraction — SQL from chart definitions is parsed for table references

Query Log Mining​

For data warehouses, Qarion mines query history directly from audit logs:

  • Snowflake — QUERY_HISTORY views
  • BigQuery — JOBS information schema
  • PostgreSQL — pg_stat_statements

This produces aggregate metrics: query count, unique users, last queried date, and popular columns.

Governance and Risk Signals​

Usage Analytics also incorporates stewardship context:

  • Product owners and stewards
  • Access requests and active subscriptions
  • Recent incidents linked to products
  • Stale active grants that have not been used within the selected window
  • Product type, domain, environment, and criticality metadata

Key Features​

Dependency Scores​

Every product receives a dependency score (0-100) based on adoption and risk signals. Searches, views, queries, access requests, subscriptions, and incidents increase the score; ignored signals and stale grants reduce it. The score helps stewards prioritize products that people depend on or that need attention.

Usage Analytics Page​

The Usage Analytics page ranks products in the current space by dependency score, queries, views, incidents, or last activity. You can filter by:

  • Time window: 7, 30, 90, 180, or 365 days
  • Product type
  • Owner or steward
  • Data domain
  • Criticality

Selecting a product opens a detail drawer with signal counts, recent incidents, suggested actions, and a shortcut to the product's Usage tab.

Dashboard Adoption Widget​

The dashboard includes a compact usage adoption widget for day-to-day review. It shows depended-on products, recurring incident products, ignored or stale products, and the highest-priority product rows for the selected window.

Usage Tab​

The Usage tab on each product's detail page shows:

  • Query frequency over time
  • Top querying users
  • Popular columns (most referenced in WHERE/GROUP BY clauses)
  • Common query patterns (anonymized)

Unused Products View​

Products with zero queries in the last N days are surfaced as candidates for deprecation, helping you keep the catalog clean and trustworthy.

Dashboard Lineage​

Scraped dashboards appear as nodes in the lineage graph, with edges connecting tables to the dashboards that consume them. Each dashboard node displays the BI tool icon and a link back to the original dashboard.

Getting Started​