electricsheepafrica/african-port-throughput-performance

---
license: cc-by-4.0
tags:
- trade
- ports
- logistics
- shipping
- sub-saharan-africa
- synthetic
- supply-chain
- maritime
- infrastructure
- africa
pretty_name: African Port Throughput Performance
size_categories:
- 10K<n<100K
task_categories:
- tabular-classification
- tabular-regression
---

# African Port Throughput Performance

Synthetic dataset modelling container throughput and operational performance for 15 major African seaports across three scenarios.

## Scenarios

| Scenario | Description |
|---|---|
| `baseline` | Current operational conditions |
| `capacity_expansion` | Investment in cranes, berths, and digital customs |
| `congestion_crisis` | Surge in demand with infrastructure bottlenecks |

Each scenario contains **10 000 records** covering 2019–2025.

## Ports

Mombasa (Kenya), Dar es Salaam (Tanzania), Lagos/Apapa (Nigeria), Durban (South Africa), Tema (Ghana), Djibouti (Djibouti), Maputo (Mozambique), Lome (Togo), Abidjan (Côte d'Ivoire), Dakar (Senegal), Walvis Bay (Namibia), Beira (Mozambique), Douala (Cameroon), Pointe Noire (Republic of Congo), Cape Town (South Africa).

## Variables

| Column | Type | Description |
|---|---|---|
| `record_id` | string | Unique identifier (UUID) |
| `port_name` | string | Port name |
| `country` | string | Country |
| `year` | int | Calendar year (2019–2025) |
| `quarter` | string | Quarter (Q1–Q4) |
| `container_throughput_teu` | int | Twenty-foot equivalent units |
| `vessel_turnaround_hours` | float | Hours from arrival to departure |
| `berth_occupancy_rate` | float | 0–1 |
| `crane_moves_per_hour` | float | Container moves per crane-hour |
| `dwell_time_days` | float | Average container dwell time |
| `cost_per_container_usd` | float | USD per TEU handled |
| `connectivity_index` | float | 0–100 liner connectivity score |
| `customs_integration_score` | float | 0–100 digital customs score |
| `congestion_index` | float | 0–1 |
| `port_efficiency_class` | string | efficient / moderate / inefficient / congested |

## Generation

```bash
pip install -r requirements.txt
python generate_dataset.py --output-dir data --records-per-scenario 10000
python validate_dataset.py data
```

## License

CC BY 4.0