WoAG Occupation Coding Service User Guide

The Whole-of-Australian-Government (WoAG) Occupation Coding Service (‘the Coding Service’ or ‘the service’) has been designed by the Australian Bureau of Statistics (ABS) to provide a single occupation coder across government, industry and the community. 

The Coding Service will code occupation data to the latest Australian standard occupation classification titles and codes.

The Coding Service has been built with supervised machine-learning technology, to train hierarchical support vector machine (HSVM) models that provide high-quality and comprehensive automated coding against hierarchical classification categories. A confidence threshold is applied to the service, ensuring that outputs are high quality.

The service is called via an Application Programming Interface (API), designed to support integration across systems and platforms (including online forms and survey instruments) by offering authenticated, standards-compliant endpoints. 

All API services are hosted in Australia to comply with relevant data sovereignty and privacy regulations. 

Users have the option to register as a public user (throttled service), or a partner user (enhanced capability). Public and partner registrations enable the following services:

• Single record (synchronous) coding
• Small batch synchronous coding (up to 300 records)

• Single record (synchronous) coding
• Small batch synchronous coding (up to 300 records)
• Large-file asynchronous upload/download bulk coding (from 1 record to millions of records)

Public user real time coding API calls will be throttled at 1 request per second, within a ceiling of 1000 requests per day. Limits on partner service usage may also apply, subject to review.

The Coding Service and API have been security assessed by an independent registered assessor within the Australian Signals Directorate (ASD) Information Security Registered Assessors Program (IRAP) Program. This assessment found the Coding Service and API to have met the control and security objectives defined through the Australian Government Information Security Manual (ISM). 

The service has been built to comply to the ISM and the Protective Security Policy Framework (PSPF). It leverages modern web API technologies in accordance with the Australian Government’s API Standard and globally recognised security frameworks. These standards ensure that the service is designed for safe, scalable integration across government and public-facing systems. 

Both public and partner service users will be registered, and will be provided with relevant authorisation tokens to access the service.

See Coding Service security for more detail, including security controls to assist partner agencies in assessing their risks when using this service.

This user guide supports access to and use of the WoAG Occupation Coding Service.  It is targeted toward software developers and technical professionals integrating the service into a client application. 

The guide outlines the API endpoints available for accessing and using the service, and provides integration instructions for calling the API. It is structured to be followed sequentially from Getting started through to Gathering parameters. 

Users will then proceed to Real-time (synchronous) coding for single record or small batch coding, or Asynchronous batch coding, depending on the data to be coded. 

• The synchronous single-record coding service is designed for real-time usage (~1 second per record). It is suitable for small volumes and live systems such as online forms and web surveys. Synchronous coding also supports coding small batches of records (up to 300 records) with similar per-record timing.
• Note: This service is not optimised for large volumes and should not be used for high-throughput workloads. Use asynchronous coding for scalable batch processing.

• Asynchronous batch coding is designed for large datasets (from a single record to millions of records). While asynchronous coding is the most efficient service for larger batches of data, it is not real-time, and may be queued during high load periods.
• Batch uploads are submitted via the API, and status is checked via polling (operation endpoints).
• Response times for batch requests may range from a few minutes to several hours depending on file size, system demand and current queue load at the time of submission.