View the live dashboard here:
https://ai-logistics-command-center.streamlit.app
An executive-grade logistics intelligence system that simulates how an e-commerce company can monitor fulfillment risk, warehouse utilization, courier performance, inventory exposure, priority exceptions, and revenue at risk.
This portfolio project was designed for companies operating high-volume logistics environments such as Amazon-style fulfillment networks, Shopee, Lazada, courier operations, retail distribution, and supply chain control towers.
The AI Warehouse & Logistics Command Center analyzes 500 simulated e-commerce orders and automatically identifies operational risks across warehouse routing, courier performance, inventory availability, customer tier exposure, SLA pressure, and revenue-at-risk impact.
The system converts raw logistics data into management-ready insights:
- Which orders are likely to be delayed
- Which warehouse should fulfill each order
- Which courier should be assigned
- Which exceptions require escalation
- Which teams own the operational issue
- How much revenue is exposed
- Which regions, warehouses, and couriers are driving risk
- What leadership should do next
This project demonstrates how AI, automation, Python, FastAPI, and Streamlit can be used to create a logistics control tower for proactive exception management.
E-commerce and logistics teams often operate reactively.
Common issues include:
- Late deliveries discovered too late
- Warehouse congestion not visible early enough
- Courier delay patterns not escalated quickly
- Inventory risks hidden across fulfillment locations
- High-value customer orders treated the same as low-risk orders
- Manual exception tracking across spreadsheets and emails
- Limited executive visibility into revenue at risk
The result is poor customer experience, missed SLAs, revenue leakage, avoidable escalations, and operational firefighting.
This project creates a logistics intelligence layer that analyzes every order and produces:
- Warehouse recommendation
- Courier recommendation
- Inventory stockout risk
- Delay risk score
- Exception classification
- Priority level
- Owning team
- Recommended action
- Executive KPIs
- Dashboard-ready CSV and JSON outputs
- Streamlit management dashboard
- FastAPI endpoints for automation and integration
The project includes a Python data generator that creates a realistic logistics dataset:
- 500 customer orders
- 500 delivery records
- 6 warehouses
- 8 courier partners
- 12 product SKUs
- 72 warehouse-SKU inventory records
Generated files:
app/data/orders.csv
app/data/delivery_history.csv
app/data/warehouses.csv
app/data/couriers.csv
app/data/inventory.csv
The inventory engine classifies SKU-warehouse inventory positions as:
- Low risk
- Medium risk
- High risk
- Critical risk
It also identifies replenishment status:
- Healthy
- Monitor Closely
- Reorder Needed
- Critical Replenishment Needed
This helps operations teams identify which items need immediate replenishment or closer monitoring.
The warehouse recommendation engine scores available fulfillment locations using:
- Customer region match
- Product availability
- Safety stock impact
- Warehouse utilization
- Daily load pressure
- Stockout risk
It recommends the best warehouse and explains the routing decision.
The courier recommendation engine scores courier partners using:
- Customer region coverage
- On-time rate
- Delay rate
- Average delivery days
- Cost per kilometer
- Delivery promise window
It recommends the best courier and explains the courier selection logic.
Each order receives a delay risk score and category:
- Low
- Medium
- High
- Critical
The score considers:
- Warehouse congestion
- Courier delay rate
- Courier risk level
- Warehouse selection risk
- Inventory risk
- Tight promised delivery windows
- Customer tier
- Regional routing mismatch
The output includes a recommended action for operations teams.
The exception classifier identifies the main operational issue behind each risky order.
Exception types include:
- Courier Delay
- Stockout Risk
- Warehouse Congestion
- SLA Breach Risk
- Inventory Shortage
- High Value Customer Risk
- Manual Review Required
- No Immediate Exception
Each exception is assigned:
- Priority level: P1, P2, P3, or P4
- Owning team
- Escalation requirement
- Recommended action
The KPI engine calculates leadership-level metrics such as:
- Total orders
- Active exceptions
- Active exception rate
- P1/P2 priority exceptions
- Revenue at risk
- Predicted delay rate
- OTIF estimate
- Perfect order rate estimate
- Escalation rate
- Platinum and Gold customer orders at risk
- Courier delay count
- Stockout risk count
- Warehouse congestion count
The export engine generates dashboard-ready output files:
app/outputs/logistics_kpis.csv
app/outputs/order_risk_report.csv
app/outputs/exception_type_summary.csv
app/outputs/owning_team_summary.csv
app/outputs/warehouse_performance_summary.csv
app/outputs/courier_performance_summary.csv
app/outputs/region_summary.csv
app/outputs/executive_summary_input.json
These outputs can be used by:
- Streamlit
- n8n
- Claude
- Power BI
- Tableau
- Looker Studio
- Other reporting tools
The project includes a FastAPI backend with endpoints for accessing logistics intelligence.
Important endpoints:
GET /health
GET /orders
GET /warehouses
GET /inventory
GET /couriers
GET /delivery-history
GET /analyze/orders
GET /exceptions
GET /exceptions/priority
GET /kpis
GET /summaries/exception-types
GET /summaries/owning-teams
GET /summaries/warehouses
GET /summaries/couriers
GET /summaries/regions
GET /export
POST /order/analyze
The API can be used by automation tools such as n8n to trigger alerts, summaries, and workflow actions.
The Streamlit dashboard includes:
- Executive Overview
- Order Risk Monitor
- Warehouse Performance
- Courier Performance
- Inventory Risk
- AI Executive Brief
The dashboard is designed for management users and includes director-level interpretation, operational insights, and recommended action plans.
Based on the generated dataset, the system produced the following management signals:
Total Orders: 500
Overall Status: At Risk
Active Exceptions: 269
P1/P2 Priority Exceptions: 129
Revenue at Risk: ₱4,382,716
OTIF Estimate: 74.20%
Predicted Delay Rate: 25.80%
Perfect Order Rate Estimate: 46.20%
Main risk signals:
Leading actionable issue: Courier Delay
Affected courier delay orders: 169
Stockout risk records: 53
Warehouse congestion records: 8
Most exposed region: Mindanao
Raw Logistics Data
↓
Data Loader
↓
Inventory Risk Engine
↓
Warehouse Selection Engine
↓
Courier Selection Engine
↓
Delay Risk Prediction Engine
↓
Exception Classification Engine
↓
Executive KPI Engine
↓
Export Engine
↓
FastAPI + Streamlit Dashboard + n8n/Claude-ready JSON
Python
pandas
FastAPI
Uvicorn
Streamlit
Plotly
pytest
n8n-ready JSON outputs
Claude-ready executive summary input
logistics-command-center/
├── app/
│ ├── data/
│ ├── models/
│ ├── outputs/
│ ├── services/
│ │ ├── data_loader.py
│ │ ├── inventory_service.py
│ │ ├── warehouse_selection_service.py
│ │ ├── courier_selection_service.py
│ │ ├── delay_prediction_service.py
│ │ ├── exception_classifier_service.py
│ │ ├── kpi_service.py
│ │ └── export_service.py
│ ├── config.py
│ └── main.py
├── dashboard/
│ └── streamlit_app.py
├── docs/
│ └── screenshots/
├── n8n/
├── scripts/
│ └── generate_mock_data.py
├── tests/
├── requirements.txt
├── README.md
└── .gitignore
git clone <your-repository-url>
cd logistics-command-centerpython3 -m venv .venv
source .venv/bin/activatepip install -r requirements.txtpython3 scripts/generate_mock_data.pypython - <<'PY'
from app.services.data_loader import DataLoader
from app.services.export_service import run_full_export
loader = DataLoader()
result = run_full_export(
loader.load_orders(),
loader.load_warehouses(),
loader.load_inventory(),
loader.load_couriers()
)
print(result["message"])
PYuvicorn app.main:app --reloadOpen:
http://127.0.0.1:8000/docs
In a separate terminal:
source .venv/bin/activate
streamlit run dashboard/streamlit_app.pyOpen:
http://localhost:8501
Endpoint:
POST /order/analyze
Sample payload:
{
"order_id": "ORD-9999",
"customer_id": "CUST-9999",
"customer_location": "Makati City",
"customer_region": "Metro Manila",
"product_sku": "SKU-CAMERA-006",
"quantity": 1,
"order_value": 39000,
"customer_tier": "Platinum",
"order_date": "2026-07-01",
"promised_delivery_date": "2026-07-02"
}The API returns warehouse recommendation, courier recommendation, delay risk score, exception type, priority level, owning team, and recommended action.
This project demonstrates:
- Business process improvement thinking
- Logistics and operations analytics
- AI transformation project design
- Python data pipeline development
- FastAPI backend development
- Streamlit executive dashboard development
- Exception management logic
- KPI and revenue-at-risk storytelling
- n8n/Claude-ready automation architecture
It is designed to show how automation and AI can move a logistics team from reactive firefighting to proactive exception management.
Planned improvements:
- n8n workflow for P1/P2 exception alerts
- Claude-generated executive brief
- Email or Slack notification automation
- Machine learning delay prediction model
- Scenario simulation for courier rerouting
- Inventory replenishment recommendation engine
- User authentication
- Live database integration
- Cloud deployment
Built as a portfolio project demonstrating AI, automation, logistics analytics, process improvement, and executive dashboarding capability.