Zia Security Events — 1B Synthetic Records
Dataset Description
A production-grade dataset containing over 1,030,000,000 rows of high-fidelity enterprise security logs. Built specifically for training ML threat detection models, stress-testing SIEM/SOC pipelines, and benchmarking big data applications — without exposing real credentials or PII.
- Massive Scale: 1B+ rows pre-sharded into 10 Parquet files (~78.3 GB)
- Rich Schema: 10 core columns including IPs, timestamps, risk scores, and SHA-256 hashes
- Realistic Distribution: Covers 12 security categories and 36 actions with a true-to-life 85% Success / 15% Failure split
Instant Free Sample
Test the quality immediately. No account or signup required.
👉 Download 50-Row Free Sample (CSV)
Try the AI Challenge
Paste this free sample into Gemini or ChatGPT. Both models independently flag the data as a live enterprise security breach — voluntarily generating Splunk rules, MITRE ATT&CK mappings, and threat-hunting reports.
Access & Pricing
1. Full Dataset Access — $50.00
All 1.03 Billion rows on Hugging Face.
- Pay securely via PayPal: Click Here to Pay
- Email your Hugging Face username to: zia.data.team@protonmail.com
- Your account will be whitelisted within 24 hours.
2. Custom 1 Billion Row Dataset — $499.99
Built to your exact specifications. Contact zia.data.team@protonmail.com for details.
Need enterprise-scale licensing (28B+ rows)? Contact zia.data.team@protonmail.com
How to Load
from datasets import load_dataset
# Load the dataset
ds = load_dataset("ziadatalabs/Zia-Security-Events-1B-Synthetic")
# Print the first row
print(ds['train'][0])
Data Schema & Fields
| Field Name | Data Type | Description |
|---|---|---|
| id | int64 | Unique row ID |
| timestamp | int64 | Unix epoch timestamp |
| admin_id | string | Synthetic administrator identifier |
| client_ip | string | RFC-1918 synthetic IP address |
| category | string | Security category (e.g., DLP, Malware Protection) |
| action | string | Action taken (e.g., Block, Quarantine) |
| result | string | Outcome (Success / Failure) |
| score | float64 | Risk score (0.0 - 1.0) |
| source | string | Data provenance |
| sha256_hash | string | Per-row integrity hash |
Technical Specifications
- Format: Apache Parquet (10 Shards)
- Total Records: 1,030,000,000
- Total Size: ~78.3 GB
- License: Proprietary / Custom Commercial (Internal R&D and Training)
- Producer: Zia Data Labs (2026)
Strictly Prohibited: Public redistribution, resale, or mirroring of the raw Parquet shards is forbidden under our commercial terms.
Citation
Zia Data Labs, 2026 https://huggingface.co/datasets/ziadatalabs/Zia-Security-Events-1B-Synthetic
- Downloads last month
- 262