The Australian Taxation Office (ATO) is taking its first steps into using data analytics to improve tax audits and track down tax evaders, with a call for applications for companies to join a new data analytics multi-use list.
The ATO will pick up services, products, and hardware through the list for the agency's Smarter Data Program, outlined in the ATO's latest corporate plan released last year.
"We will better use technology and data analytics to identify and deal with those seeking not to comply," said the ATO corporate plan document, released in June. "For most non-compliant taxpayers, it will be a matter of 'when' we catch up with them, not 'if'."
The companies that sign onto the list would be invited to go through procurement processes and propose solutions for specific issues identified under the Smarter Data Program.
The ATO is establishing a sandpit known as the RAD Lab (Research and Development Lab) that allows agile testing of products and processes before they are deployed into the ATO's production environment.
Inside this lab, the ATO is already testing an Hadoop-based Enterprise Data Hub consisting of five virtualised servers running Ubuntu, 12 physical Hadoop worker nodes running Red Hat, and three management nodes.
Some of the use cases provided by the ATO include allowing the agency to compare a person's declared income and expenses to their known assets and earnings relative to their peer groups, perform risk assessments, perform industry benchmarking, and behaviour prediction such as a person's willingness to comply based on historical data such as credit ratings.
The agency would also like to track use of its websites, with data including click logs, content, layouts, and where users may be struggling to find information.
There is no limit on the number of companies that can be added to the list. The ATO has said it is open to using open-source products and cloud-based products, and may use the whole-of-government Cloud Services Panel being established by the Department of Finance this month.
The list is expected to be established by the end of March, running until March 2020 at the latest.