Skip to content

Live demo

← Back to docs home

Below is the actual dashboard, pre-loaded with the synthetic sample data. It runs entirely in your browser — no server, no upload, no install. The same dashboard ships with the pipeline release and works the same way against your own real bundles.

What you can try

  • Browse Patients (default) -- "Jane Demo" appears in the left sidebar. Click her to load the per-patient view, then click through the Docs, Visits, Problems, Meds, Labs/Obs, and Sections tabs.
  • Keyword Search -- toggle the mode in the sidebar header. With this rich bundle there are lots of hits to find: try riluzole, edaravone, dysphagia, BiPAP, PEG, pseudobulbar, palliative, or any LOINC code like 8480-6. Each hit shows a snippet from where the term appears, with the match highlighted, and clicking a hit jumps to the relevant patient + tab.
  • Format filters on Docs -- filter to CCDA, then notice the same document is reachable as a section narrative in the Sections tab.
  • Export buttons -- "Export CSV" on any tab downloads a CSV with patient name, MRN, and DOB columns prepended. "Export hits" in keyword-search mode does the same for every search match.

What you're seeing

The demo bundle was generated by running the actual pipeline against the sample CCDA + FHIR data documented under Sample data. The bundle has 84 records spread across every category the dashboard supports:

  • 1 patient -- Jane Marie Demo, female, born 1965-03-15
  • 4 encounters -- initial neurology consult, MDC clinic visits, an inpatient hospitalization for PEG placement
  • 6 problems -- ALS plus disease-related comorbidities (dysphagia, respiratory weakness, dysarthria, fatigue, depression), each with both SNOMED and ICD-10 codes
  • 8 medications -- disease-modifying (riluzole, edaravone) and symptom-management (baclofen, dextromethorphan-quinidine, trazodone, tizanidine, vitamin D3, lorazepam), with RxNorm codes, dosing, route, and frequency
  • 4 procedures -- EMG, MRI brain, BiPAP initiation, PEG tube placement
  • 17 observations -- 7 vital signs (including FVC trending down) and 10 lab values, all with units and reference ranges
  • 3 allergies -- penicillin (high severity), sulfa, latex
  • 4 immunizations -- flu, COVID booster, PCV13, Tdap, with lot numbers
  • 3 care plans -- multidisciplinary, palliative, communication
  • 3 diagnostic reports -- EMG findings, MRI read, pulmonary function
  • 3 goals -- functional independence, symptom management, ACP
  • 9 CCDA section narratives -- the human-readable text from each populated CCDA section

Compared to a real run, the proportions differ -- a typical export has hundreds to thousands of patients with tens of thousands of documents. The dashboard's pagination, format filters, and global search are sized for that scale.

How the embedding works

The dashboard's main script first checks for a window.EMBEDDED_BUNDLE variable. If present, it loads that data and hides the file picker. Otherwise the file picker shows up as it normally does in the standalone dashboard.

The full embedded-demo HTML is at assets/dashboard_demo.html -- about 65 KB including the dashboard code (~60 KB) and the embedded JSON bundle (~9 KB).

For your own deployments, you can either:

  • Use the standard dashboard. Ship dashboard.html plus a dashboard_data.json and have users load the JSON via the file picker. This is the lowest-friction option and what most teams use.
  • Embed your own bundle. Take any dashboard.html, add a <script>window.EMBEDDED_BUNDLE = { ... }</script> block before the main script tag, save as a single self-contained HTML file. Useful when sharing a one-off cohort snapshot with a collaborator who doesn't need the full pipeline.