Data scientist guide

Tax for Data Scientists Australia

This page is for data scientists who want a clearer starting point on common deductions, including cloud compute costs, software subscriptions, conferences, working-from-home expenses, GPU hardware, professional memberships, and mixed-use device records.

Quick answer: data scientists can often claim a range of work-related expenses, but the ATO requires a direct connection to earning your income and you cannot claim anything your employer reimbursed. The main traps are private use of hardware and cloud services, overclaiming home office expenses, and failing to keep mixed-use device records.

Common data scientist deductions

Often deductible

  • Cloud compute costs (AWS, GCP, Azure) used for work projects and model training
  • Software subscriptions (data tools, IDEs, statistical packages, notebook platforms)
  • Conference and training fees that maintain or improve current data science skills
  • Home office running costs using the ATO fixed-rate or actual-cost method
  • GPU hardware, laptops, monitors, and peripherals used for work
  • Professional memberships and industry association fees
  • Technical books and online learning platform subscriptions used for work

Often non-deductible

  • Private use portion of hardware, cloud services, and software
  • Normal home-to-office commuting
  • Cloud compute or GPU costs for personal side projects unrelated to earning income
  • Hardware or software supplied or reimbursed by your employer
  • Courses aimed at a completely different profession
  • General clothing and personal expenses

Cloud costs, hardware, and mixed-use checkpoints

  • Cloud compute: subscription and usage-based costs are deductible in the year incurred. Apportion between work and personal use where the service is used for both.
  • GPU hardware: items costing $300 or less can be claimed immediately; items over $300 are depreciated over their effective life. Mixed-use devices require records of work vs private usage.
  • Home office: the ATO offers a fixed-rate method (67 cents per hour) or an actual-cost method. Keep records of hours worked from home.
  • Mixed-use records: for devices used for both work and personal purposes, keep a diary or log of usage to support the work-related percentage claimed.
  • Reimbursements: if your employer paid you back, you generally cannot claim the same cost.

Records data scientists should keep

  • Invoices for cloud compute services, software subscriptions, and data tools
  • Purchase records for hardware with depreciation schedules for items over $300
  • Usage logs or diary records for mixed-use devices (work vs personal split)
  • Home office usage diary or timesheet showing hours worked from home
  • Conference receipts, training invoices, and related travel records
  • Professional membership receipts
  • Evidence of any employer reimbursement to avoid double-claiming

Start with these calculators

Data scientist tax FAQs

Can data scientists claim cloud compute costs?

Generally yes where the cloud services are used for work. If also used for personal projects, only the work portion is deductible. Keep invoices and usage records.

Can data scientists claim GPU hardware?

Usually yes. Items under $300 can be claimed immediately; items over $300 are depreciated. Keep records of work vs private usage for mixed-use devices.

Can data scientists claim conference and training costs?

Generally yes where the content maintains or improves skills used in your current data science role. Courses aimed at a different profession are usually not deductible.

Tax Accuracy & Sources

Reviewed: March 2026 · Tax year: 2025-26

This guide summarises common data scientist deduction patterns only. Always check whether the expense was reimbursed, whether any private element needs apportionment, and whether the cloud service or hardware has the connection the ATO requires.

Uses 2025-26 ATO rates.