electricsheepafrica/african-manufacturing-pmi-data

---
license: cc-by-4.0
language:
- en
tags:
- manufacturing
- pmi
- economic-indicators
- sub-saharan-africa
- synthetic
- macroeconomics
task_categories:
- tabular-regression
- time-series-forecasting
pretty_name: African Manufacturing Purchasing Managers' Index (PMI) Data
size_categories:
- 100K<n<1M
---

# African Manufacturing PMI Dataset

Synthetic Purchasing Managers' Index (PMI) data for eight African manufacturing economies across three macroeconomic scenarios.

## Dataset Description

This dataset contains **240,000 synthetic records** of manufacturing PMI data calibrated to realistic distributions observed in Sub-Saharan African and North African economies. Each record represents a monthly observation of composite PMI and sub-indices for a specific country, sector, and scenario.

### Countries

| Country | Region | Key Manufacturing Sub-sectors |
|---------|--------|-------------------------------|
| South Africa | Southern Africa | Automotive, Metals, Food Processing, Chemicals |
| Nigeria | West Africa | Food Processing, Textiles, Cement, Chemicals |
| Kenya | East Africa | Food Processing, Textiles, Chemicals, Metals |
| Ethiopia | East Africa | Textiles, Leather, Food Processing, Chemicals |
| Egypt | North Africa | Food Processing, Metals, Chemicals, Nonmetallic Minerals |
| Morocco | North Africa | Automotive, Textiles, Aerospace, Food Processing |
| Ghana | West Africa | Food Processing, Textiles, Chemicals, Metals |
| Tanzania | East Africa | Food Processing, Textiles, Chemicals, Metals |

### Scenarios

- **Baseline** (80,000 records): Normal economic conditions, no systematic shift
- **Expansion** (80,000 records): Positive demand shock (+2.5 PMI points, higher volatility)
- **Contraction** (80,000 records): Economic downturn (-3.0 PMI points, higher volatility)

### Variables

| Variable | Description | Range |
|----------|-------------|-------|
| `record_id` | Unique identifier | - |
| `country` | Country name | 8 countries |
| `year` | Observation year | 2015–2025 |
| `month` | Observation month | 1–12 |
| `pmi_composite` | Composite PMI index | 40–62 |
| `output_index` | Manufacturing output | 35–65 |
| `new_orders_index` | New orders received | 35–65 |
| `employment_index` | Employment levels | 35–62 |
| `supplier_delivery_times_index` | Supplier delivery performance | 38–62 |
| `input_prices_index` | Input cost pressures | 40–75 |
| `output_prices_index` | Output price changes | 38–68 |
| `inventory_index` | Inventory levels | 35–62 |
| `export_orders_index` | Export order volumes | 30–65 |
| `business_confidence_index` | Business sentiment | 30–75 |
| `sector` | Manufacturing sub-sector | - |
| `pmi_signal` | Expansionary classification | expanding / contracting / stable |

### PMI Signal Thresholds

- **Expanding**: PMI > 50.5
- **Stable**: 49.5 ≤ PMI ≤ 50.5
- **Contracting**: PMI < 49.5

### Country PMI Calibration

PMI distributions are calibrated to approximate real-world patterns:
- South Africa: mean ≈ 50–52 (established industrial base, higher volatility)
- Nigeria: mean ≈ 48–51 (structural headwinds, currency pressures)
- Kenya, Ethiopia, Tanzania: mean ≈ 50–52 (growing industrialization)
- Egypt, Morocco: mean ≈ 49–52 (mixed industrial maturity)

## Generation

```bash
pip install -r requirements.txt
python generate_dataset.py
python validate_dataset.py
```

## Intended Use

- Macroeconomic analysis of African manufacturing trends
- Training ML models for economic time-series forecasting
- Scenario analysis and stress testing
- PMI calibration benchmarking

## Limitations

- This is a **synthetic** dataset. It does not represent actual PMI survey data.
- Correlations between sub-indices are approximated, not derived from real panel data.
- Seasonal patterns are not explicitly modeled.
- Inter-country dependencies (e.g., trade linkages) are not captured.

## License

CC-BY-4.0

## Citation

```bibtex
@dataset{african_manufacturing_pmi_2025,
title={African Manufacturing Purchasing Managers' Index (PMI) Dataset},
year={2025},
publisher={Electric Sheep Africa},
license={CC-BY-4.0}
}
```