DataLens

DataLens originated as the internal business intelligence platform inside Yandex before its release as an independent analytics suite. The system pushes queries directly to underlying SQL engines such as ClickHouse and PostgreSQL rather than extracting records into a proprietary memory cube. This allows operations teams and agencies to build multi-tenant dashboards over live warehouse data without running specialized desktop software.

Why DataLens instead of Power BI

Power BI centers on the Microsoft ecosystem, relying on Windows-only Power BI Desktop for data modeling, DAX for expressions, and Azure for cloud sharing. Connecting Power BI to columnar stores like ClickHouse often requires ODBC bridges or gateway agents that add latency. DataLens takes a warehouse-first approach: it connects over native database protocols to ClickHouse, PostgreSQL, MySQL, and cloud analytical stores, translating dataset metrics into source-native SQL queries or letting engineers write parameterized SQL in QL charts.

The integration model shifts where transformations occur. Power BI expects teams to perform ETL in Power Query and cache datasets on Microsoft infrastructure; DataLens leaves data at rest in the warehouse, eliminating sync jobs and per-viewer Pro or Premium seat fees. What you forfeit is Power BI's ecosystem of pre-built commercial connectors, automated Excel export pipelines, and deep DAX time-intelligence functions. For teams standardizing on SQL engines rather than the Microsoft stack, DataLens removes desktop dependencies and vendor telemetry.

What it does well

• Native ClickHouse and SQL pushdown: executes aggregations directly in the source database, returning sub-second query results without duplicating data into an external cache.
• QL charts for direct SQL queries: lets engineers and analysts write raw, parameterized SQL queries that feed dashboard widgets directly, bypassing dataset builders.
• Workbooks and collections hierarchy: organizes connections, datasets, and dashboards into folder workspaces with viewer, editor, and admin access roles.
• Parameterized dashboard selectors: links drop-down filters, date pickers, and text inputs across charts through shared dataset fields or SQL variables.
• Containerized deployment: packages backend query workers, a Node.js metadata layer, and a web UI into Docker Compose and Helm templates for self-hosted infrastructure.