Step 3 — Atemporal / Temporal Dashboard

cohort_graph.json → Mode A (Gestalt) / Mode B (Trajectory) toggle (Section 4) · runs entirely in-browser, no Python/Node required · same file feeds Step 4 next
What this does: loads the Step 2 output (cohort_graph.json) and renders the same patient graph two ways: Mode A (Atemporal/Gestalt) strips timestamps and clusters each patient's full observation/diagnosis/plan set into one profile view; Mode B (Temporal trajectory) renders the timestamped edges as a timeline. One toggle, one rendering path, for both T1D and T2D — node styling (spike markers, severity sizing, comorbidity role) is driven entirely by generic node attributes, not by disease-specific branches in the code.

Output / handoff: this stage is read-only — it does not transform or re-save the graph. The same cohort_graph.json you loaded here carries forward unchanged as the input to Step 4 (pivot lens selector), which re-anchors this identical node/edge set around Disease, Intervention, or a Disease + Intervention combo — no re-ingestion or schema change needed. There is no export button on this page because there is nothing new to export yet.
No file loaded yet — export cohort_graph.json from Step 2 first.