Dashboard Interfaces
Relevant source files
The following files were used as context for generating this wiki page:
Dashboard Interfaces in ZINC Fusion V15 provide institutional-grade visualization and decision-support tools for soybean oil (ZL) procurement. The system transitions from a high-level landing page to specialized views covering primary forecasting, procurement strategy, legislative impacts, and sentiment analysis. These interfaces are designed to translate complex multi-layer ensemble ML outputs into actionable insights, such as procurement "postures" (e.g., WAIT, ACCUMULATE) and risk-adjusted price targets.
Sources: Docs/zinc-fusion-v15-preview.html:1000-1050, frontend/src/app/page.tsx:10-30, frontend/src/app/strategy/page.tsx:15-35
Primary Dashboard Architecture
The main dashboard is structured as a high-density grid featuring real-time price tracking, multi-horizon forecast cones, and the "Specialist Ensemble" signal status. It utilizes a status bar for global state visibility, including the current ZL price, market regime (e.g., STABLE, VOLATILE), and model confidence levels.
Components and Layout
- Status Bar: Fixed-position header displaying real-time ZL price delta, regime chips (Crisis, Stable, Rising, Falling), and high-level confidence badges.
- Chart Area: Features ZL Soybean Oil front-month futures with overlayed forecast cones representing P30–P70 (primary) and P10–P90 (outer risk) intervals.
- Key Risk Indicators (Big 4 Gauges): Visualizes the most critical drivers: VIX Stress, Harvest Pace, China Tension, and Tariff Threat.
- Sidebar: Contains the "Procurement Posture" directive and the detailed Specialist Ensemble (12 domain-specific models).
Sources: Docs/zinc-fusion-v15-preview.html:30-150, Docs/zinc-fusion-v15-preview.html:850-950
The diagram represents the structural relationship between UI elements and the underlying intelligence modules displayed on the dashboard.
Sources: Docs/zinc-fusion-v15-preview.html:850-1100
Procurement Strategy Interface
The Strategy Page provides decision-support tools for physical procurement timing. It centers around a "Current Posture" HUD that translates ML conviction into human-readable directives like "ACCUMULATE" or "WAIT."
Key Strategic Modules
- Primary Directive: A prioritized list of actions (e.g., "Cover Q1 Needs," "Defer H2") with specific volume recommendations.
- Factor Waterfall: Decomposes ZL price changes into specific driver contributions (e.g., Technical Momentum vs. China Demand).
- Scenario Analysis: Displays impact matrices for binary risks, such as Section 301 tariff imposition or Brazil harvest delays.
- Monte Carlo Summary: Summarizes 10,000 simulations to provide P10 (floor) and P90 (ceiling) targets for hedging.
Sources: frontend/src/app/strategy/page.tsx:15-100, Docs/zinc-fusion-v15-preview.html:1150-1250
Scenario Impact Table
| Scenario | Probability | 1M Impact | Primary Specialists |
|---|---|---|---|
| Tariff Imposed (25%) | 35% | +$4.80 | Tariff, China |
| Tariff Delayed | 45% | +$0.40 | Tariff, China, Trump |
| Brazil Harvest Delay | 25% | +$3.60 | Crush, Volatility |
| Palm Oil Glut | 20% | -$2.80 | Palm, Substitutes |
Sources: Docs/zinc-fusion-v15-preview.html:1250-1320
Specialist Ensemble Visuals
The Specialist Ensemble represents 11-12 isolated domain models that feed into the core unified forecast. Each specialist has a dedicated UI card or row showing its directional signal (-1.0 to +1.0) and weight contribution to the ensemble.
| Specialist | Scope | Key Inputs |
|---|---|---|
| Crush Economics | Processing margins | ZL/ZS/ZM spreads, NOPA |
| China Demand | Import flows | GACC, MOFCOM, CNY |
| Energy Complex | Feedstock correlations | Crude, BOHO ratio, Diesel |
| Tariff Regime | Trade policy | Section 301, Retaliatory duties |
| Biofuel Mandate | Policy demand | RFS, RINs, SAF |
| Volatility | Risk regimes | VIX, OVX, CVOL |
Sources: Docs/zinc-fusion-v15-preview.html:550-750, frontend/src/app/page.tsx:50-150, grafana/scripts/generate_specialist_dashboards.py:10-50
This sequence shows the data flow from specialized domain models through the core ensemble to the final UI rendering.
Sources: frontend/src/app/page.tsx:10-30, grafana/scripts/generate_specialist_dashboards.py:200-250
AI-Powered Synthesis and Sentiment
The interfaces utilize Large Language Models (specifically Claude Opus 4.5) to generate natural language synthesis of quantitative data. This feature is integrated into the "Market Intelligence" cards found on the dashboard and sentiment pages.
- Sentiment Feed: Displays FinBERT-scored news with specialist routing (e.g., a news item tagged for "Biofuel" and "Energy").
- Anti-Bullshit Gate: A technical requirement where AI-generated reports MUST echo the
asOfDatefrom the input data to verify freshness and prevent hallucinations. - Synthesis Output: Includes TL;DR summaries, technical outlooks, and "Invalidation Triggers" (conditions that would change the current model view).
Sources: frontend/src/lib/ai-intelligence.ts:10-100, Docs/zinc-fusion-v15-preview.html:1500-1600
AI Market Intelligence Structure
{
"headline": "10 words max summarizing ZL outlook",
"reasoning": "2-3 sentences explaining cross-driver dynamics",
"zlOutlook": "BULLISH | NEUTRAL | CAUTIOUS | BEARISH",
"keyRisks": ["risk 1", "risk 2"],
"comprehensiveReport": {
"tldr": "Quick summary with price targets",
"keyDrivers": "Breakdown of Biofuel, Weather, Macro, Trade",
"forecasts": "Time-horizon percentage moves"
}
}
Sources: frontend/src/lib/ai-intelligence.ts:80-110
Conclusion
ZINC Fusion V15 Dashboard Interfaces act as the "Intelligence Command Center" for US Oil Solutions. By combining high-fidelity charts, risk-assessment gauges, and AI-synthesized market reports, the system provides a comprehensive bridge between complex quantitative ML pipelines and practical procurement strategy. The multi-page SPA architecture ensures that users can drill down from aggregate market pressure scores to individual specialist performance and sentiment drivers.
Sources: Docs/zinc-fusion-v15-preview.html:1800-1850, frontend/src/app/page.tsx:180-200