Hugging Face's logo

Datasets:
electricsheepafrica
/
nigerian-banking-web-banking-sessions

Tasks:
Tabular Classification
Modalities:
Tabular
Text
Formats:
parquet
Languages:
English
Size:
1M - 10M
Tags:
nigeria
banking
finance
fraud-detection
Synthetic
african-data
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files
xet
Community
nigerian-banking-web-banking-sessions / README.md
Kossisoroyce's picture
Kossisoroyce
Upload README.md with huggingface_hub
e733f1a verified
preview code
|
raw
history blame contribute delete
2.42 kB
metadata
license: apache-2.0
task_categories:
  - tabular-classification
tags:
  - nigeria
  - banking
  - finance
  - fraud-detection
  - synthetic
  - african-data
language:
  - en
size_categories:
  - 1M<n<10M

Web Banking Sessions

Dataset Description

Nigerian web banking sessions with session hijacking fraud labels

This is a production-grade synthetic dataset with authentic Nigerian banking context, designed for web fraud detection, phishing detection, and browser-based threat monitoring.

Key Features

  • 🇳🇬 Nigerian Context: Real bank names, states, demographics
  • 📊 ML-Ready Labels: Fraud flag (0.8% prevalence)
  • ⚡ Large Scale: 3,000,000 rows
  • 🎯 Realistic Distributions: Lognormal session durations, browser distributions, hourly patterns
  • 📅 Temporal Coverage: 2023-01-01 to 2024-12-31

Dataset Structure

Splits

  • Train: 2,100,000 rows (70%)
  • Validation: 450,000 rows (15%)
  • Test: 450,000 rows (15%)

Schema

Key columns: session_id, session_start/end, customer_id, account_id, device_id, browser, browser_version, session_duration_seconds, actions_count, transactions_count, login_method, state, ip_address, session_status, fraud_flag

Usage 

from datasets import load_dataset

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

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

Labels

  • Type: Fraud flag (0.8% prevalence)
  • Use Case: web fraud detection, phishing detection, and browser-based threat monitoring

Nigerian Context

Nigerian web banking patterns, browser distribution (Chrome 55%, Safari 20%, Firefox 15%), typical session behaviors, and fraud patterns including credential stuffing and session replay.

Citation 

@dataset{nigerian_banking_web_banking_sessions,
  title = {Nigerian Web Banking Sessions},
  author = {Electric Sheep Africa},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/electricsheepafrica/nigerian_web_banking_sessions}}
}

License

Apache 2.0 - Free for commercial and research use.

Contact