electricsheepafrica/nigerian-banking-agent-network

Agent Network (Super Agents, Master Agents, Sub-Agents)

Dataset Description

Nigerian banking agents and their transactions with fraud labels

This is a production-grade synthetic dataset with authentic Nigerian banking context, designed for agent fraud detection, network optimization, and commission management.

Key Features

• 🇳🇬 Nigerian Context: Real bank names, states, demographics
• 📊 ML-Ready Labels: Fraud flag (2.5% prevalence)
• ⚡ Large Scale: 2,000,000 rows
• 🎯 Realistic Distributions: Lognormal volumes, agent type hierarchies, commission structures
• 📅 Temporal Coverage: 2023-01-01 to 2024-12-31

Dataset Structure

Splits

• Train: 1,400,000 rows (70%)
• Validation: 300,000 rows (15%)
• Test: 300,000 rows (15%)

Schema

Key columns: agent_id, agent_name, state, agent_type (super/master/sub), monthly_volume_ngn, monthly_transactions, commission_rate, kyc_tier, fraud_flag

Usage

from datasets import load_dataset

# Load full dataset
dataset = load_dataset("electricsheepafrica/nigerian_agent_network")

# Load specific split
train_data = load_dataset("electricsheepafrica/nigerian_agent_network", split="train")

Labels

• Type: Fraud flag (2.5% prevalence)
• Use Case: agent fraud detection, network optimization, and commission management

Nigerian Context

Nigerian agent banking network structure, super agent chains, rural agent distributions across 37 states, KYC tier distributions, and fraud patterns including float diversion and fake transactions.

Citation

@dataset{nigerian_banking_agent_network,
  title = {Nigerian Agent Network (Super Agents, Master Agents, Sub-Agents)},
  author = {Electric Sheep Africa},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/electricsheepafrica/nigerian_agent_network}}
}

License

Apache 2.0 - Free for commercial and research use.

Contact

• Organization: Electric Sheep Africa
• Collection: Nigeria Banking & Finance Datasets