Apache Superset vs Metabase
After Databricks acquired Redash, Metabase and Apache Superset were the last two popular open-source business intelligence platforms left that were open source.
In this post, we'll showcase where Metabase shines and struggles in comparison to Preset Cloud / Superset.
Preset's founder and CEO is Max Beauchemin, the original creator of Apache Superset. At Preset, we're the experts of Superset and we're proud to be the leading vendor offering Apache Superset as a cloud hosted service. In addition to all of the features that Superset has out of the box, we add the following:
- Preset Manager: our admin layer that lets you create multiple Superset workspaces and manage permissions using role-based access control (RBAC)
- Documentation & Training: we maintain high quality end-user focused documentation and also offer training
- Compliance & Security: Preset is SOC2 Type 1 & Type 2 compliant and we offer single sign-on (SSO) for paid teams
- World-class Support: teams on a paid tier get access to our dedicated email and chat support
The Superset community enjoys the benefit of being a project housed in the Apache Software Foundation. The barrier to contributing is low and the Superset community has added support for new databases at a breakneck pace.
|NoSQL databases||1 (MongoDB)||0 (possible through Trino / Presto)|
|SQL Engines||3 (Presto, Databricks, Athena)||5 (Presto, Trino, Drill, Athena, Databricks)|
|Support common SQL operations (Select, Filter, GroupBy, etc)||✅||✅|
|Preview visualization during iteration||✅||✅|
|Preview generated SQL query during iteration||✅||✅|
|View data transformation results as a table||✅||✅|
|Use virtual datasets to build charts||✅||✅|
|Ability to join tables||✅||➕ In Development|
Both Metabase and Superset have no-code interfaces for crafting SQL queries.
Business intelligence tools are only as good as the visualizations they enable.
|Common charts (pie, line, bar, etc)||✅||✅|
|Clear documentation on adding your own visualizations||❌||✅|
|Number of default charts||17||40 (and growing 📈)|
|Number of geospatial visualizations||3||10|
Similar to how the Superset community helped drive support for over 20 databases, they've also been the driving force behind the large array of visualizations in Superset. This is no surprise, since Superset's contributor community is significantly larger than Metabase's.
|Export query results as CSV||✅||✅|
|Database metadata explorer||✅||✅|
|Linting and auto-complete||✅||✅|
|Support for variables in SQL queries||✅||✅|
|Save query for re-use||✅||✅|
|Semantic layer: save queries as virtual datasets||✅||✅|
|Semantic layer: define custom metrics||✅||✅|
|Semantic layer: calculated columns||✅||✅|
Both Metabase and Superset ship with a powerful SQL editor and a lightweight semantic layer.
- Aggregate values across multiple columns and publish as Metrics.
- Metrics can be certified as authoritative by a specific user.
- Transform specific columns and publish as Calculated Columns.
- Write arbitrary SQL queries and publish as a Virtual Dataset.
In the no-code chart builder (Explore), metrics, columns, and virtual datasets all inherit the power that physical database tables have in Superset.
|Basic datetime, value, and range filters||✅|