Faasera Documentation

Faasera User Guide: Synthetic Data Generation

Faasera’s Synthetic Data Engine enables the generation of realistic, lineage-aware datasets for non-production environments. It supports multiple generation modes, schema preservation, referential integrity, and seeded randomness — all aligned with compliance and privacy goals.

Key Features

Feature Description
Schema-Based Generation Generates data based on existing schema or user-defined templates
Seeded Randomness Ensures repeatability for test data generation using fixed seeds
Referential Integrity Maintains relationships between foreign keys and primary keys
Context-Aware Fields Applies domain logic (e.g., age vs. DOB, state vs. zip)
Privacy-Enhanced Defaults Avoids re-identification by generating safe, non-real data
AI-Augmented Value Suggestions Optionally uses AI hints for plausible names, locations, or values

Configuration Modes

You can define generation rules per table and per column using a JSON policy structure. Key elements include:

{
  "tableName": "customers",
  "generationType": "SYNTHETIC",
  "columnRuleMap": {
    "firstname": { "type": "SEED", "seedFileName": "firstnames.csv" },
    "lastname": { "type": "SEED", "seedFileName": "lastnames.csv" },
    "email": { "type": "PRESERVE", "pattern": "{firstname}.{lastname}@example.com" },
    "dob": { "type": "DATE_RANDOM", "range": { "start": "1980-01-01", "end": "2000-12-31" } }
  },
  "primaryKeyColumn": "customer_id",
  "compositeKeyColumn": "customer_id+dob"
}

Generation Types

Type Description
SEED Picks value from a seeded list (e.g., names, cities)
RANDOM Random value with optional format (e.g., strings, numbers)
PRESERVE Derived based on another field (e.g., email from firstname/lastname)
DATE_RANDOM Generates random date within a specified range
FPE_ENCODED Applies format-preserving encoding using FPE keys

Referential Integrity

To preserve key relationships:

Using AI for Plausible Generation (Optional)

You can enrich values using external hinting or AI copilots (e.g., to generate more localized names, job titles, etc.)

Output & Integration

Example Use Cases

Use Case Benefit
QA/Test Data Generation Enables safe, representative test environments
Pre-Sales Demos Creates realistic demo datasets
Data Science Bootstrapping Generates data for ML model experimentation
Sandboxing New Features Safe testbed without real PII/PHI

Compliance Notes

Need Help?