+
{{ value }}
+{{ hint }}
++ Relative sensor positions in the 580 m cluster. Not a city-wide heatmap. +
+
+ Uploads through Nitro to Go /api/train. The control credentials stay in the Node process and are
+ never sent to the browser.
+
+ {{ errorMessage }} +
+ ++ {{ job.error }} +
+About · method, data and limits
++ PulseCast turns a footfall forecast into a next-week staffing plan. The forecast is a model output; the + staffing rule is a transparent heuristic. Both are shown, editable, and bounded. +
++ There is no POS, sales, queue, payroll or labour-cost feed. Cost is an assumption. The coverage score + is high by construction because shift blocks cover their local peak; overstaffed hours are the honest + trade-off. +
+Meta comes from Go /api/meta through the Nitro proxy.
Presenter control · not linked in public navigation
+
+ Each button POSTs to this Nuxt origin. Nitro attaches the Go control credentials server-side; the browser
+ never sees them.
+ The enum is fixed by the API contract. Weather and event values match the demo narrative. +
+
+ Put /stage on the projector and let the audience scan the QR code. /live needs no
+ login and keeps the last known plan if the stream drops.
+
Forecast · {{ forecastWindow }}
+
+ A 168-hour forecast for {{ activeSensor?.name ?? 'the selected store' }}, aggregated by
+ day. The decision, shift blocks and hourly coverage are on
+ Peak estimated devices per hour. Weather is context, not a causal claim. +
++ Scroll the table sideways on a phone for the uncertainty band and staffing columns. Weekly mean + {{ formatNumber(total) }} estimated devices/hour across open+closed hours. Staff at peak uses 25 + visitors/staff-hour × 1.15; the plan can use a different standard. +
++ The weather window does not cover the plan week. +
++ Weather is the main accuracy driver in the model. Holidays are context and a smaller feature. +
+Now · replay clock {{ formatDateTime(demoNow) }}
+
+ Eight fixed sensors around Stadtgalerie and Kaiserstraße. This is estimated device density, not people,
+ and not a city-wide heatmap. The decision lives on
Last available hour in the API series.
+{{ lastPoint ? formatNumber(lastPoint.v) : '—' }}
+estimated devices / hour
++ {{ activeSensor?.name }} · {{ activeSensor?.id }} +
++ This is a replay snapshot, not a live sensor feed. The replay clock is + {{ formatDateTime(demoNow) }}; the last measured point is + {{ lastPoint ? formatDateTime(lastPoint.t) : '—' }}. +
++ Select a sensor to update the reading. Positions are relative; this is not a geographic map. +
+Measured points, stable order by ID. Battery may be null in the API.
++ Coverage {{ formatDate(meta.coverage.from) }} → {{ formatDate(meta.coverage.to) }} · + {{ meta.coverage.cadence_min }}-min cadence · data {{ meta.data_version }} · model {{ meta.model_version }} +
+Live · no login
++ {{ activeSensor?.name ?? 'Stadtgalerie' }} · {{ scenarioLabel(live.scenarioId) }}. Updates arrive over SSE; + if the stream drops, this page polls every 5 seconds. +
++ {{ peakStaff }} staff + No plan yet +
+
+
+ {{ formatWeekday(peakShift.date) }} {{ formatShortDate(peakShift.date) }}
+ {{ peakShift.start }}–{{ peakShift.end }} · peak shift
+
+ Waiting for the peak shift
+
+
Staff-hours
+{{ plan ? formatNumber(totalStaffHours) : '—' }}
+Labour cost
+{{ plan ? formatEuro(plan.summary.cost_eur) : '—' }}
+Coverage
+{{ plan ? formatPercent(plan.summary.coverage_pct, 1) : '—' }}
+Understaffed
+{{ plan ? `${plan.summary.understaffed_hours} h` : '—' }}
+{{ live.error }}
+Plan · {{ weekLabel(weekStart, weekEnd) }}
++ Forecast for {{ activeSensor?.name ?? apiPlan.sensor ?? 'the selected store' }} needs + {{ peakHour?.required_staff ?? '—' }} staff at the busiest hour. The rest of the week + runs on {{ totalStaffHours }} staff-hours at {{ formatEuro(plan.summary.cost_eur) }}. +
++ Mock mode · NUXT_PUBLIC_USE_MOCK=1 · no Go API call. + + Local preview · the Go API plan is unchanged. + + + Go /api/staffing · scenario v{{ live.version }} · broadcast + {{ live.scenarioAt ? formatDateTime(live.scenarioAt) : '—' }} · no POS, sales or labour-cost feed. + +
+Peak staffing decision
+{{ peakShift?.staff ?? '—' }} staff
+
+
+ {{ formatWeekday(peakShift.date) }} {{ formatShortDate(peakShift.date) }}, {{ peakShift.start }}–{{ peakShift.end }}
+
+
+ {{ formatWeekday(berlinParts(peakHour.t).date) }} {{ String(berlinParts(peakHour.t).hour).padStart(2, '0') }}:00
+ peak: {{ peakHour.footfall }} est. visitors/hour
+
+
+ Store, service standard and scenario are sent to Go /api/staffing. Local assumption edits render a + preview from the API forecast. +
+Service standard
+{{ standardLabel }}
++ These four values are not query inputs of Go /api/staffing. Changing them shows a local preview; + “Use API plan” restores the server values. +
+2–5 h blocks per day; staff count covers the block peak.
++ {{ formatWeekday(day.date) }} {{ formatShortDate(day.date) }} +
++ Total {{ totalStaffHours }} staff-hours · {{ formatEuro(plan.summary.cost_eur) }} at + €{{ plan.assumptions.labor_cost_eur_h }}/h · exported CSV has one row per shift block. +
++ {{ selectedCovered }}/{{ selectedHours.length }} open hours covered · {{ selectedOver }} over requirement +
++ Scheduled staff is constant inside a block. Coverage is at or above the service standard by + construction; the trade-off is visible as overstaffed hours. Store hours 10:00–20:00 local. +
+Simulate · {{ weekLabel(weekStart, weekEnd) }}
++ A scenario changes forecast footfall, not history. Compare the shift requirement against the base week + for {{ activeSensor?.name ?? 'Stadtgalerie' }}. +
+The contract enum is fixed; the API applies the perturbation.
+Peak required staff per day, base vs scenario.
++ {{ activeSensor?.name ?? 'Stadtgalerie' }} · {{ scenarioLabel(live.scenarioId) }} · + {{ plan ? `${totalStaffHours} staff-hours` : 'waiting for the API' }} · updates live via SSE +
+
+
{{ liveUrl }}
+{{ plan ? peakStaff : '—' }}
++ {{ formatWeekday(peakShift.date) }} {{ peakShift.start }}–{{ peakShift.end }} + Per shift block +
+Labour cost
+{{ plan ? formatEuro(plan.summary.cost_eur) : '—' }}
+{{ plan ? `€${plan.assumptions.labor_cost_eur_h}/h assumption` : '—' }}
+Coverage
+{{ plan ? formatPercent(plan.summary.coverage_pct, 1) : '—' }}
++ {{ plan ? `${plan.summary.understaffed_hours} h under` : '—' }} +
+Forecast peak
+{{ forecastPeak ? Math.round(forecastPeak) : '—' }}
+estimated devices/hour
+vs flat schedule
+−{{ plan ? formatEuro(saving) : '—' }}
+same coverage, fewer hours
++ + Put {{ peakShift.staff }} staff on {{ formatWeekday(peakShift.date) }} + {{ formatShortDate(peakShift.date) }}, {{ peakShift.start }}–{{ peakShift.end }}. + + Pick a store and scenario; the recommendation appears here. +
+
+
+ Forecast refresh failed: {{ forecastError.statusMessage }} +
+