SDVThe Synthetic Data VaultEnterprise
Built with enterprise data in mind
Demand precision from your synthetic data models.
Use a variety of transformers
With our novel synthetic data types (sdtypes) transform any data, phone numbers, emails, addresses, URLs and PII.
Learn MoreUse contextual anonymization
Extract context from complex data type and model correlations between context and resynthesize.
Learn MoreModel complex graphs
Model data consisting of arbitrary number of tables with complex relational structures.
Learn MoreGenerate conformative data
Get data that conforms to ranges, preserves referential integrity, cardinality.
Learn More
Source Available
Enterprise
Single Table
Sequential Data
Multi-sequence data
Single-sequence forecasting
Multi Table
≤ 5 tables and ≤ 10 columns, depth=1
Unlimited tables with unlimited depth and columns
Data Creation
Must learn from real data
Learn from real data or generate from scratch
Anonymization
Pure and pseudo-anonymization
Anonymization specially for: Phone numbers, lat/lang addresses, emails, URLs, IP and more
Modeling Capabilities
Basic data shapes (discrete and continuous) with missing values
Advanced capabilities: Outliers, imbalanced data, special values, and more
Constraints
9 predefined constraints, with custom constraints
Unlimited predefined constraints or custom constraints
Domain-specific settings
Insurance, Healthcare, Fintech
Synthetic Data Evaluation
Quality and Diagnostic reports, Basic metrics for ML efficacy, Privacy
Custom metrics for software testing, Data augmentation, Data portability
Benchmarking
Run benchmark manually on public models
Monthly benchmark reports with private repository of models
SDV apps
Use programmable data stack
Access pre-programmed apps
