Four paths from laboratory records to a dashboard.
Choose a familiar export, import a file into Catalyst, or query a connected source. These paths serve different needs; they are not equivalent competitors.
OpenELIS reporting · Recorded demonstrations
Choose a starting point
Start with a report you already know, or ask a new question. These pathways serve different needs: they are complementary choices, not equivalent alternatives to rank. OpenELIS reporting works independently of Catalyst and AI.
Path 1 · Export and visualize2:33
Use a familiar OpenELIS report
OpenELIS report builder → CSV → Superset
Choose fields and filters, generate a CSV, then upload it into Superset to build a table, chart and dashboard. No Catalyst or AI step is required.
Data: a snapshot for the selected reporting period. Export and upload again when you need a new snapshot.
Upload the report, review its columns and types, and save a Dataset. Create charts, arrange a dashboard and publish it to Superset without writing SQL.
Data: an imported snapshot with a saved version. A later file does not silently replace an existing published report.
Import it into Catalyst and review the detected columns and types.
Save and reopen the imported Dataset without using SQL.
Create a table and count chart, arrange a dashboard and publish it to Superset.
Path 3 · Ask the database1:56
Explore OpenELIS directly
OpenELIS PostgreSQL → Catalyst → Superset
Browse available data, ask and refine a question, review the proposed SQL and choose when to run it. Save the result as a Dataset, then build and publish a dashboard.
Data: queried from the connected operational database. This avoids a file export; results depend on the query and when it runs.
Recorded outcome: context preparation stopped before either selected model ran. The demonstration continues through Catalyst's supported manual-SQL recovery path.
Select the FHIR-derived Spark source and write a reporting question.
Review the generated and reviewed SQL before execution.
Retrieve the result and make an explicit analyst refinement.
Save the Dataset and publish its dashboard to Superset.
All four recordings were reviewed at normal speed. These are real local workflows with synthetic records, recorded September 16, 2026. The linked server dashboards separately verify the updated deployment; the recordings are not presented as server-video evidence. Owner acceptance remains open. The applications currently use their existing sign-ins; shared sign-in and equivalent per-user access are separately planned.
Captions remain visible for reading, result views are held for inspection and accelerated waits are labelled. Final owner acceptance remains a separate milestone.