Processing…
Supply Chain Dashboard
Real-time risk monitoring ·
Live Intelligence Dashboard
Real-time supply chain risk overview across all active nodes
Overall Risk Score
62
▲ +4.2% from last week
🎯
Active Suppliers
247
▲ 12 new this month
🏭
Avg Delivery Delay
8.3d
▼ improving vs last month
🚛
Inventory Health
74%
▲ Above safe threshold
📦
⚠️
Weather Disruption Detected
Severe storms expected in Southeast Asia this week. Suppliers B and D may face 5–10 day delays. Recommend pre-ordering critical components.
📈
Risk Score Trend (30 Days)
Live
🥧
Risk Distribution
🏭
Supplier Performance
📊
Feature Metrics (30d avg)
AI Risk Predictor
Configure supply chain parameters and get an instant ML-powered risk assessment
🎛
Supply Chain Parameters
Historical on-time delivery rate (0 = unreliable, 100 = perfect)
Current demand intensity relative to capacity (0 = low, 100 = extreme)
Current or forecasted shipment delay in days (0–30)
Severity 0–10
% of target stock
🔮

Configure parameters and click "Run AI Risk Prediction" to see your analysis

Random Forest · 300 estimators
Alerts & Warnings
Automated risk alerts based on threshold breaches and ML predictions
🔔
Active Alerts
🚨
Critical: Supplier C — Reliability Collapse
Supplier C on-time rate has dropped to 38% over the last 14 days. Consider immediate qualification of backup vendors. Estimated impact: 12-day production delay.
🌪️
High: Logistics Route #7 — Force Majeure Risk
Typhoon Ketsana approaching primary sea lanes. 78% probability of 8+ day delay for all shipments via Taiwan Strait. Reroute recommended via Indian Ocean.
📈
Warning: Demand Spike Detected — SKU Group B
Demand for electronics components exceeded forecast by 34% this week. Current inventory buffer will last approximately 11 days at this rate.
🏭
Warning: Port Congestion — Rotterdam Hub
Rotterdam port experiencing 5–7 day congestion. 23 active shipments affected. Consider air freight for critical orders.
✅
Info: Supplier A — Performance Improvement
Supplier A on-time delivery rate improved to 94% (up from 87% last month). Risk score reduced from Medium to Low.
Prediction History
Log of all recent ML risk predictions this session
🕒
Recent Predictions
ID Timestamp Risk Level Confidence Supplier Rel. Transport Delay Weather
No predictions yet. Run a prediction to see history.
Model Insights
Understanding how the ML model works and what drives predictions

🔄 Retrain Model

Generate a fresh 5,000-sample dataset and retrain the Random Forest classifier in-place.

Test Accuracy
—%
🎯
CV Mean ± Std
—
📊
Training Samples
4K
🗄
Predictions Logged
0
🔮
🧬
Feature Importances
Loading…
📖
How the Model Works

Algorithm: Random Forest Classifier with 300 decision trees. Each tree votes on the risk category (Low / Medium / High), and the majority vote wins. Probabilities are computed via soft voting across all trees.

Training Data: 5,000 synthetic records generated using domain-expert heuristics to simulate real supply chain behaviour. Features are scaled via StandardScaler before training.

Features chosen cover the five primary drivers of supply chain disruption: supplier reliability (historical performance), demand level (strain on capacity), transport delay (logistics bottlenecks), weather impact (external shocks), and inventory level (buffer availability).

Future improvements: Integrate real-time ERP data feeds, add LSTM time-series forecasting, incorporate geopolitical risk indices, enable multi-supplier network graph analysis, and add explainability via SHAP values per prediction.