ZBY Synthetic Data

Synthetic-data technology for complex relational, linked and regulated data.

ZBY develops practical tools, public datasets and evaluation resources for synthetic data beyond isolated flat files.

Our work focuses on:

• relational and multi-table data;
• longitudinal and cross-table consistency;
• practical utility for real analytical workloads;
• privacy-aware validation;
• provenance and controlled distribution;
• local and on-premises deployment.

Public Resources

ZBY Linked Healthcare Reference Dataset

A fully fictional, ZBY-generated linked dataset for public demonstration and evaluation.

View the dataset

The dataset contains no real patient records or restricted source data and is not intended to reproduce or represent any official healthcare dataset.

Relational Synthetic Data Evaluator

A browser-local evaluator for comparing compatible relational synthetic-data submissions with an Emulate-generated reference baseline.

Evaluation runs locally in the browser. Submitted files are not uploaded to an Emulate server.

Open the evaluator

Why This Matters

Synthetic data can appear plausible at column or table level while still failing when used as a connected dataset.

Relational quality also depends on whether records remain coherent across tables, across time and across the analytical questions the data is expected to support.

Passing basic structural checks alone does not demonstrate useful relational synthetic data.

Emulate

Emulate is building an interoperable synthetic-data workflow for complex data environments.

The public resources hosted here demonstrate selected capabilities using fully fictional data. They do not represent the complete scope of the Emulate platform.

Areas of Interest

Healthcare · Insurance · Pharmaceutical research · Relational databases · Privacy-enhancing technologies · Data quality · Data governance · On-premises AI