Analytics overview

Turn governed data models into reusable, secure, warehouse-neutral analytics and focused read-only shares.

Last verified 2026-08-28

The analytics side of ReportPlane turns stored records into controlled questions your product can ask repeatedly.

Instead of exposing a warehouse or accepting arbitrary SQL, you publish an Analytics API. The definition states what may be grouped, measured, filtered, ordered, and returned.

The analytics path

Data model → Analytics API → Query client → /query → Warehouse → Normalized result
                         ↘ Analytics Share → Unlisted read-only chart

The Analytics API selects model fields and calculation rules. A query client limits which published definitions a consumer may run. The runtime validates request-time filters and executes a warehouse-neutral plan.

For a focused human-facing result, an Analytics Share can expose one enabled, published Analytics API through an unlisted link with optional password protection. It uses the same runtime contract and does not create a separate query or data copy.

What a definition contains

  • dimensions for grouping or labels;
  • metrics for calculations;
  • allowed runtime filters and per-field operators;
  • optional fixed policy filters;
  • time buckets;
  • ordering and maximum rows;
  • chart type and renderer-neutral visualization metadata;
  • a destination binding and cache policy.

Build without a second query system

The Analytics Explorer, saved Analytics APIs, preview screens, and public query runtime share the same analytics planner and executors. Explorer is a workspace over the real runtime, not a separate implementation with different results.

Warehouse-neutral execution

Supported shapes run across BigQuery, ClickHouse, Snowflake, and Elasticsearch/OpenSearch. Multi-dimension grouping, multiple metrics, supported time buckets, grouped order, and limit follow a shared contract.

Results

Every successful public query returns:

  • columns, describing output names and value types;
  • rows, containing normalized result objects;
  • result_metadata, describing preferred x, y, and series roles plus compatible chart suggestions.
Was this page helpful?Send feedback