ServicesAboutNotesContact Get in touch →
EN FR
Note

Elementary data quality dashboards

Hub for building data quality dashboards with Elementary: generating reports, hosting them for team access, building custom BI dashboards, and designing KPIs.

Planted
elementarydbtdata qualityanalytics

Elementary stores every test result in the warehouse as queryable tables, enabling historical visibility into data quality trends — not just whether tests pass in the current run. This hub covers the progression from generating a first report to building custom KPIs in a BI tool.

The progression

Start with the generated report. One command — edr report — produces a self-contained HTML file with test history, anomaly charts, model runtime trends, and lineage. No server required. If this is enough for your team, stop there. See Elementary HTML report generation for the flags that matter and Elementary report sections for what to look at in each part of the report.

Host it so the team has access. An HTML file on your laptop doesn’t help anyone else. Elementary’s edr send-report command uploads the report to S3, GCS, or Azure Blob Storage and serves it as a static website. Automate it as a post-run step in your CI pipeline and the report stays current. See Elementary report hosting.

Build custom dashboards when you need more control. The HTML report covers most needs, but when you want executive-level views with your own branding, unified visibility across multiple dbt projects, or data quality metrics alongside operational metrics in your existing BI tool, query Elementary’s warehouse tables directly. Any tool that connects to your warehouse works. See Elementary custom BI dashboards for the key tables and example SQL.

Design KPIs that translate test results into meaningful metrics. Raw pass/fail counts are operational; KPIs are strategic. Test pass rate, SLA compliance, test coverage, and anomaly detection rate give stakeholders a way to understand data quality health without knowing how dbt tests work. See Data quality KPIs from Elementary.

Organize dashboards for the audience, not for completeness. A single all-models report with no filtering is technically comprehensive and practically overwhelming. Filtering by domain, criticality tier, and refresh cadence creates dashboards people actually use. See Elementary dashboard organization.

What this assumes

This progression assumes Elementary is already installed and running in your dbt project. If it’s not, the Elementary for dbt note covers installation, the two-component architecture (dbt package plus edr CLI), and the anomaly detection tests that feed into these dashboards.