Preview Synthetic Data

Preview Synthetic Data contains artificially generated data based on the structure and characteristics of selected WA Health datasets. It does not contain information about real people or real events. 

Preview Synthetic Data is:

• Synthetic – Contains artificially generated data and does not represent real people or real events
• Non-Representative – Does not reproduce real-world statistics or counts
• Univariate – Individual variables are designed to reflect similar patterns and value distributions to the source data, but relationships between variables are not preserved. Combinations of variables and records should not be interpreted as representing real-world patterns.

These characteristics mean Preview Synthetic Data can help applicants understand what data is available and how it is structured, but it cannot be used to understand what occurs in the real population.

What Preview Synthetic Data can be used for

Preview Synthetic Data may be used for early planning and preparation as part of an application for real data, and is limited to the following defined use cases:

Univariate Data Exploration

• Familiarisation with the names, structure, format, quality and available variables of a dataset.
• Refining an application for real data following receipt of the Preview Synthetic Data.
• Adapt research or project plans to ensure alignment with the structure, quality, and limitations of the requested data, prior to finalising data access.
• Experimentation and data visualisation with synthetic datasets.

Development and Preparation

• Prepare and refine data pipelines or scripts for processing real data (e.g. data ingestion, cleaning, transformation and validation).
• Develop and verify analytical code and workflows prior to accessing real data.
• Develop and refine dashboards, reports and visualisations using synthetic data.

What Preview Synthetic Data can not be used for 

Preview Synthetic Data is intended for exploration and preparation only. It must not be used:

• For formal statistical analysis, decision-making, or publication
• To produce findings or draw conclusions
• To inform operational, clinical, or policy decisions
• For reporting on individuals, individual conditions or individual healthcare establishments
• As a replacement for real linked health data

Preview Synthetic Data supports exploration and preparation only, and is not a substitute for real data. All substantive analysis and validation must be conducted using approved real data.

Two Preview Synthetic Data products are currently available.

20-row Preview Synthetic Data Samples

Small, non-representative synthetic data samples that are publicly available through the data collections page.

These samples:

• Contain 20-rows of non-representative, univariate synthetic data
• Replicate the structure, variables and format of real datasets
• Do not contain real or identifiable information
• Support data discovery and early familiarisation

These samples allow applicants to understand what data is available and how it is structured before applying for real data.

1-year Preview Synthetic Datasets

Larger, non-representative synthetic datasets available to applicants as part of a real data application.

These datasets:

• Replicate the structure and variables of real data across approximately one year 
• Do not contain real or identifiable information
• Support preliminary testing and workflow preparation

Preview Synthetic Data products currently include:

Accessing the 20-row Preview Synthetic Data Samples

The 20-row synthetic samples are publicly available through the data collections page. No application for real data is required to access these small samples.

Accessing the 1-year Preview Synthetic Datasets

To access a 1-year Preview Synthetic Dataset, applicants must:

• lodge an application for real data through ISPD Client Services
• complete the Synthetic Data Access Request Form (external link)
• agree to the Synthetic Data Information Release Agreement through the Synthetic Data Access Request Form.

Applicants can only access Preview Synthetic Data for the data collections included in their real data application. 

Access to Preview Synthetic Data is provided to support preparation for the real data application and does not constitute approval to access or use real data.

The 20-row synthetic samples are publicly available and may be used for early data exploration and visibility across a dataset's properties. Access to the 1-year synthetic dataset samples is restricted to defined use cases and must be used strongly in accordance with the approved use case.

Applicants must comply with the Synthetic Data Information Release Agreement presented within the Synthetic Data Access Request form, which outlines user obligations, security requirements, and conditions for handling synthetic data.

Applicants are responsible for ensuring that Preview Synthetic Data is used appropriately and in line with all applicable conditions and limitations.

Access to Preview Synthetic Data does not alter or replace existing governance arrangements for real data. All access to real data remains subject to standard ethics, Data Custodian, and approval processes.

All WA health entity users of Preview Synthetic Data must comply with the WA Health Aboriginal Data Governance Policy and complete the relevant training. Any enquiries relating to this policy may be directed to aboriginal.health@health.wa.gov.au.

Preview Synthetic Data may be used to support early planning and preparation, but all substantive analysis, validation, and decision‑making must be undertaken using approved real data.

Further detail on synthetic data governance and obligations is available in the Synthetic Data User Guide (PDF 361KB).

WA Health offers a range of synthetic data products designed to support different research, analytical, system testing and data discovery needs. These products vary in their level of detail, representativeness, functionality and permitted uses, with each product tailored to specific approved purposes.

To learn more about WA Health’s synthetic data capabilities and available products, visit the synthetic data page.