Dashboards

Dashboards

Planned operational dashboards (distinct from the read-only Reporting screens): a live landing page with state-aware heroes for each operational area, plus a Dashboards sidebar menu with deeper per-area views.

Spec (research-enriched, June 2026): docs/dashboardScope.md — ISA-101 two-tier display hierarchy, Stephen Few hero rules, ISA-18.2 alert-fatigue doctrine, and Dematic InSights vendor evidence; 21 sources, 22 adversarially verified claims.

Two-tier architecture

Following ISA-101 Level 1 / Level 2 practice (verified):

Tier	Location	Purpose
Level 1 — Landing page	Home (/)	At-a-glance state heroes for all five areas; top-priority alert count only; every tile is a drill-down link
Level 2 — Dashboards menu	/dashboards/*	Per-area deep dives: more metrics, history, breakdowns; still read-only

No operator actions on any dashboard screen — operators act on the operational screens (Transport, GTP, Counting). Dashboards tell operators where to go.

KPI-count discipline (verified, practitioner consensus): cap each area dashboard at 5–7 metrics; the landing page may carry up to 7–9 across its five tiles.

Landing page tiles

Five tiles. Each: primary hero + one or two secondary figures + sparkline + state colour (ok / warning / critical) + drill-down link. Defaults are configurable in Settings.

Tile	Primary hero	State trigger
Inbound	Open inbound orders + projected finish time at current receive rate	Warning: projection past end of receiving shift; Critical: receive errors today > 5 (default)
Stock blocking outbound	Lines currently unfulfillable from pickable stock (allocation shortfall)	Warning: ≥ 1 line blocked; Critical: > 5 % of today's open outbound lines blocked (default)
Outbound	Open outbound orders vs. released total + projected completion (Dematic-style)	Warning: projection past earliest unmet route cutoff; Critical: past last cutoff
Dispatch	Next route countdown + route order-readiness (ready / assigned)	Warning: projected readiness past cutoff; Critical: cutoff breached with orders not ready
Automation	Alert count by severity (shape + colour coded)	From highest active alert severity

Dashboards menu

Five screens under a "Dashboards" sidebar section — graphs and heroes only, no tables. Each caps at 5–7 metrics.

Screen	Key metrics	Status
Inbound	Time to receive (median), putaway time (median), putaway backlog + oldest age; receipts per hour vs. norm; 14-day receive-error trend	⬜
Outbound	Open orders + projected completion, lines/hour vs. target, accuracy proxy (shorts + corrections), on-time-to-cutoff rate	⬜
Replenishment	Urgent replenishments (pick-face below min with open demand), projected stockouts within 2 h, open task oldest age; urgency strip worst-first	⬜
Stock	Location utilisation drawn inside 90-day normal band; HU count, SKU count; 90-day utilisation line	⬜
ABC Movers	Pareto curve with A/B/C cut points; top-10 / bottom-10 movers as ranked horizontal bars with sparklines; risers / fallers strips	⬜

ABC ranking metric: picks (lines) per SKU over trailing 90 days (no item costs in openWCS). Default class boundaries — A: top items covering ~70–80 % of movement; B: next ~15–20 %; C: the rest — all configurable.

Hero design rules

Verified (ISA-101 + Few):

• Every hero carries a comparison (target, same period last week, norm band, or projected completion at current rate), a computed state driving its colour, and rounded values (12.4 k, not 12,407).
• Format: big number + horizontal bullet/limit bar (value against target with shaded ranges) + sparkline. Bullet bars beat radial gauges on speed and accuracy — no gauges anywhere.
• No bare trend arrows (ambiguous); a sparkline shows the same thing unambiguously.
• Multi-element panels (scanner array, conveyor segments, lanes) use aligned normalised bars or a heat strip: "all normal" reads as a flat line.
• Exception lists sort worst-first, colour only on out-of-range entries.
• Most critical hero goes top-left (most-attended region in left-to-right reading).

Colour and alert doctrine

Verified across Rockwell (ISA-101), Few, and ISA-18.2:

• Grey base theme. Live data in calm, low-saturation colours. Bright colour is reserved exclusively for abnormal states and never used decoratively. A healthy floor looks quiet.
• One alert-hue family with 2–3 intensity steps; never encode state by hue alone — pair with a shape or icon change (~8 % of males are colour-blind). Reuse the existing --warning / --danger tokens.
• Alert ≡ actionable abnormality (ISA-18.2). Anything an operator cannot act on must not surface as an alert. Thresholds are configurable defaults in Settings, not hard-coded constants.
• The landing automation tile shows an alert count by priority; the full list lives behind the drill-down.
• This intentionally diverges from the 3D twin's green/orange/red equipment skins — the twin is a spatial mimic; dashboards are KPI surfaces and follow the grey doctrine.

Refresh cadence

Default polling (configurable per deployment):

Surface	Cadence
Landing tiles	Every 15 s (same as GTP / Transport screens)
Area dashboards	Every 60 s; history charts on load

Every screen shows a last-updated stamp. If a poll fails twice, the affected tile greys out and shows "stale" rather than silently displaying old numbers.

Status

All dashboard screens are planned (⬜). The design spec and research backing are complete. The existing Reporting screens cover historical analysis; dashboards add live operational state (heroes, projections, alert counts). See Roadmap and Status — Phase 4.