What is Recce?
Recce offers contextual data impact analysis designed to streamline the pull request (PR) review process for dbt (data build tool) projects. It helps analytics engineers and PR reviewers understand the potential impact of code changes on data before merging them into production. By comparing data between development branches and the production environment, Recce provides clear visibility into how data models, structures, and values might change.
This tool facilitates faster merge times by providing proof-of-correctness for data changes, allowing stakeholders to sign off with confidence. It aims to maintain stable production data by enabling thorough quality assurance during the development cycle. Recce supports data teams in implementing best practices for PR reviews by offering actionable insights and systematic checks, ultimately improving collaboration and reducing the risk of deploying breaking changes.
Features
- Lineage Diff: Visualize and focus impact assessment on the affected area of the dbt lineage DAG.
- Structural Checks: Automated row count and schema checks to maintain data consistency.
- Data Profile Diff: Compare before-and-after data profiles using statistical comparisons for quick impact understanding.
- Low Level Checks: Examine row-level data changes for precise validation and root cause analysis.
- dbt Integration: Works with dbt artifacts (docs) from development and production environments.
- Contextual Impact Analysis: Compare development data changes against production data.
Use Cases
- Streamlining Pull Request (PR) reviews for dbt projects.
- Providing proof-of-correctness for data changes by Analytics Engineers.
- Enabling PR Reviewers to quickly assess and approve data impacts.
- Preventing breaking data changes from reaching production environments.
- Improving collaboration between developers and stakeholders on data projects.
- Implementing data quality assurance (QA) best practices within the development workflow.
- Visualizing the impact area of dbt model changes.