dbt + Materialize:
Real-Time Data Superpowers
Use the best-in-class analytics engineering workflows of dbt to manage operational data workloads on Materialize.
Keep your Standard SQL
Keep the SQL modelling conventions your team is familiar with.
End Manual Scheduling
dbt run once and your views are continually kept updated.
Get Event-Driven Capabilities
Use the push architecture of Materialize to action immediately on your data.
We decided to look into Materialize to handle personalization and feature-serving in real-time, and by that evening we were up and running.
Read Customer storyTom Cooper Head of Data, Superscript
How Materialize Integrates with dbt
What is dbt?
dbt is a SQL-based transformation workflow that lets teams quickly and collaboratively deploy analytics code following software engineering best practices.
What is Materialize?
Materialize is an operational data warehouse: The familiar architecture of an analytical data warehouse, but with a continuous streaming computation model built for workloads that need immediate action on up-to-date results.
Write SQL transformations in dbt
Maintain the entire schema of your SQL transformations (sources, views, materialized views) in a dbt project in a git repo.
Use the dbt CLI to build views in Materialize
Execute dbt run from the CLI to build the SQL models and transformations in Materialize once, and to migrate after updates.
Results are automatically kept up-to-date
Materialize automatically updates the results of SQL transformations as source data changes. Results can be queried in SQL or streamed out.
Build Operational Data Products with dbt + Materialize
Real-Time & User-Facing Analytics
→Dashboards and data products need to be reactive to up-to-the-minute changes in your business.
Automation and Alerting
→Save time for your users, and build value by taking action or notifying at only the right moments.
Segmentation and Personalization
→Value of personalization, recommendations, dynamic pricing increases as latency of data aggregations approaches zero.
ML in Production
→Online feature stores need continually updated data, operators need to monitor and react to changes in ML effectiveness.
