Skip to main content
Changelog

What is new in Golden Analytics

Release notes for every version. New features, performance improvements, and fixes.

v0.9.2 Feature Improvement

  • Databricks SQL support. Connect to Databricks SQL warehouses using a personal access token or service principal. Unity Catalog is supported natively.
  • dbt semantic layer integration. If your workspace is connected to a dbt project, Golden Analytics can resolve metric names directly from your dbt manifest. Available on Pro and Enterprise.
  • Attribution export to CSV. Download the full dimension breakdown of any attribution run as a CSV file from the results panel.
  • Query latency improvements for BigQuery connections: in our internal testing, median query time dropped 31% for tables over 50 million rows.

v0.9.0 Feature

  • Scheduled attribution scans. Configure Golden Analytics to run attribution on a metric every day at a set time. Results are stored in your attribution history and trigger alerts when anomalies are detected.
  • Slack alert integration. Send attribution summaries and anomaly alerts to a Slack channel. Configure per-metric thresholds from workspace settings.
  • Multi-warehouse workspace. Connect up to three warehouses in a single Pro workspace. Query any warehouse from the same interface.

v0.8.1 Improvement Fix

  • Improved NLQ parsing for date ranges. Natural language queries with relative date ranges ("last quarter", "year to date", "rolling 90 days") now resolve correctly across all supported locales.
  • Redshift Serverless support. Workgroup endpoint connections now work alongside provisioned cluster connections. IAM authentication is available for both.
  • Fixed: attribution results panel could show stale data when switching between saved queries in the same session.
  • Fixed: percentage change formatting showed incorrect sign for metrics with inverted polarity (lower is better).

v0.8.0 Feature

  • Shareable attribution links. Generate a read-only link to any attribution result. Recipients can view the full dimension breakdown without signing in.
  • Attribution history. All past runs are saved to a per-metric history log. Compare how driver rankings have shifted over time.
  • Snowflake Cortex metadata integration. Enrich attribution results with semantic metadata from your Snowflake Cortex definitions. Column descriptions and entity labels appear alongside dimension breakdowns.

v0.7.0 Feature Improvement

  • Natural language querying (beta). Ask attribution questions in plain English. "Why did revenue drop last week?" runs the full attribution pipeline and returns a ranked answer. Powered by your warehouse, not an external LLM.
  • Dimension filter pinning. Lock specific dimension values before running attribution to isolate a segment. Results exclude the pinned filters from the global driver ranking.
  • Schema introspection is now 4x faster on large warehouses with more than 500 tables.

v0.6.0 Feature

  • Team workspaces. Invite teammates to a shared workspace. Each user gets their own query history and saved metrics. Permissions and seat management are in workspace settings.
  • BigQuery support. Connect to any BigQuery project using a service account JSON key or Workload Identity Federation. Attribution queries run as BigQuery batch jobs using your existing reservation.
  • Saved metric library. Save frequently-used metric definitions as named entries in your workspace. Open any saved metric to run attribution instantly without re-specifying the column and grain.

v0.5.0 - Initial release Feature

  • Core metric attribution engine. Connect a Snowflake warehouse, select a metric column, define a comparison window, and receive a ranked list of dimensions explaining the change. Attribution runs directly inside your warehouse using read-only credentials.
  • Bar chart breakdown. Attribution results include a visual bar chart showing each dimension's contribution as a percentage of total change.
  • Redshift support. Connect Amazon Redshift clusters via JDBC endpoint with database user credentials.