electricsheepafrica/african-public-procurement-data

---
license: cc-by-4.0
task_categories:
- tabular-classification
- tabular-regression
language:
- en
tags:
- governance
- procurement
- anti-corruption
- sub-saharan-africa
- synthetic
- public-administration
- contracts
- transparency
pretty_name: African Public Procurement Data
size_categories:
- 10K<n<100K
configs:
- config_name: baseline
data_files: data/baseline.csv
default: true
- config_name: high_transparency
data_files: data/high_transparency.csv
- config_name: low_transparency
data_files: data/low_transparency.csv
---

# African Public Procurement Data

## Abstract

A synthetic dataset modeling public procurement contracts across 12 sub-Saharan African countries (2018–2025), parameterized from procurement authority reports, OECD assessments, and anti-corruption research. Contains 10,000 records per scenario across three transparency scenarios (baseline, high_transparency, low_transparency), with 11 variables covering procurement methods, contract values, bidder counts, single-source rates, contract completion, price benchmarking, and award timelines. Designed for ML classification, anomaly detection, and anti-corruption research in the public procurement domain.

## 1. Introduction

Public procurement accounts for approximately 12-20% of GDP across sub-Saharan African countries, representing one of the largest areas of government spending and corruption risk. Kenya alone reported over 34,000 contracts worth KES 262.8 billion in FY 2023/2024. Single-source (direct) procurement remains prevalent, with rates varying from 8% in high-transparency systems to 30% in weak governance environments.

Key challenges include: limited competition (average 2-5 bidders per tender), contract completion rates below 60% in many countries, price markups of 20-100% above benchmarks, and procurement timelines exceeding 100 days. The adoption of e-procurement systems (Kenya, Rwanda, Zambia) has improved transparency but coverage remains incomplete.

## 2. Methodology

### 2.1 Target Population

Contract-level procurement records for 12 sub-Saharan African countries spanning 2018–2025, across 10 sectors and 4 procurement methods.

**Countries included:** Nigeria, Kenya, South Africa, Ghana, Tanzania, Uganda, Rwanda, Ethiopia, Senegal, Zambia, Mozambique, Cameroon.

### 2.2 Parameterization Evidence Table

| Parameter | Value Used | Source | Year | Note |
|-----------|-----------|--------|------|------|
| Kenya contracts FY2023/24 | 34,000+ / KES 262.8B | Kenya PPRA MAPS | 2024 | E-procurement coverage |
| Single source rate (high) | ~8% | OECD benchmarks | 2024 | Open method dominant |
| Single source rate (low) | ~30% | Brookings Nigeria | 2024 | Direct procurement prevalent |
| Average bidders (open) | 4-6 | ZPPA Zambia | 2024 | Open tender competitive |
| Contract completion rate | 45-75% | Corruption Watch SA | 2024 | Varies by sector |
| Price benchmark markup | 20-100% | GTI Global PP Dataset | 2024 | Corruption indicator |
| Procurement share of GDP | 12-20% | World Bank | 2023 | SSA average |

### 2.3 Scenario Design

| Scenario | Description | Single Source Mult | Completion Mult | Bidder Mult |
|----------|-------------|-------------------|-----------------|-------------|
| **baseline** | Current SSA procurement landscape | 1.0× | 1.0× | 1.0× |
| **high_transparency** | Countries with e-procurement and reforms | 0.5× | 1.2× | 1.5× |
| **low_transparency** | Weak governance, high corruption risk | 2.0× | 0.7× | 0.6× |

## 3. Dataset Description

### 3.1 Schema

| Column | Type | Units | Range | Description |
|--------|------|-------|-------|-------------|
| record_id | int | — | 1–10,000 | Unique record identifier |
| country | categorical | — | 12 countries | Sub-Saharan African country |
| year | int | year | 2018–2025 | Procurement year |
| procurement_method | categorical | — | 4 methods | open, selective, limited, direct |
| sector | categorical | — | 10 sectors | Procurement sector |
| contract_value_usd_millions | float | USD millions | 0.01–1000+ | Contract value |
| num_bidders | int | count | 1–20 | Number of bidders |
| single_source | boolean | — | true/false | Single-source procurement flag |
| contract_completed | boolean | — | true/false | Contract completion status |
| price_benchmark_ratio | float | ratio | 0.5–3.0 | Actual price / benchmark price |
| days_to_award | int | days | 7–365 | Days from tender to award |

### 3.2 Summary Statistics (baseline)

| Variable | Mean | SD | Min | Max |
|----------|------|-----|-----|-----|
| contract_value_usd_millions | 15.2 | 45.3 | 0.01 | 850 |
| num_bidders | 3.3 | 2.1 | 1 | 15 |
| single_source rate | 0.29 | — | — | — |
| completion rate | 0.59 | — | — | — |
| price_benchmark_ratio | 1.02 | 0.20 | 0.5 | 3.0 |
| days_to_award | 85 | 45 | 7 | 365 |

## 4. Usage

### 4.1 Loading with HuggingFace datasets

```python
from datasets import load_dataset

ds = load_dataset("electricsheepafrica/african-public-procurement-data")
ds_low = load_dataset("electricsheepafrica/african-public-procurement-data", "low_transparency")
```

### 4.2 Regenerating

```bash
pip install numpy pandas scipy matplotlib
python generate_dataset.py --scenario baseline --n 10000 --seed 42
python validate_dataset.py
```

## 5. Limitations & Ethical Considerations

1. **Synthetic data**: Not suitable for audit investigations or official reporting.
2. **Country-level aggregation**: Does not capture subnational procurement variations.
3. **Sector simplification**: Procurement categories are aggregated into 10 broad sectors.
4. **No contract-level detail**: Individual contract clauses, amendments, and variations not modeled.
5. **Temporal simplification**: Does not capture fiscal year-end procurement spikes.

## 6. References

1. Fazekas et al., *Global Contract-level Public Procurement Dataset*, 2024.
2. Corruption Watch, *Procurement Watch Report 2024*.
3. Kenya PPRA, *MAPS Assessment Report 2024*.
4. Zambia ZPPA, *Procurement Statistics Reports*.
5. Brookings, *Transparency in Procurement in Nigeria*, 2024.
6. OECD, *Implementing Procurement Recommendation 2020-2024*.
7. World Bank, *Doing Business Procurement Indicators*.
8. Open Contracting Partnership, *OCDS Implementation Reports*.

## Citation

```bibtex
@dataset{esa_procurement_2026,
title={African Public Procurement Data},
author={{Electric Sheep Africa}},
year={2026},
publisher={HuggingFace},
url={https://huggingface.co/datasets/electricsheepafrica/african-public-procurement-data},
license={CC-BY-4.0}
}
```

## License

CC-BY-4.0