The risk of job losses due to automation and computerisation throughout Australia by 2030. About this data.
This interactive stems from a study by Oxford University economists Carl Frey and Michael Osborne, who in 2013 modelled the potential consequences of the expansion of automation on the US labour market.
Later, the Committee for Economic Development of Australia (CEDA) replicated and extended the study for Australia.
CEDA took Frey and Osborne’s estimates for the susceptibility of job types to automation and applied them to job codes based on the Australian and New Zealand Standard Classification of Occupations (ANZSCO).
The probability of automation per ANZSCO code was then weighted by the fraction of workers in each job type to produce an overall estimate for susceptibility of job automation in Australia.
CEDA’s job counts were taken in November 2014 at the ANZSCO ‘unit group’ level.
We’ve used CEDA’s data – published in a research report titled Australia’s future workforce – as follows:
We merged CEDA’s data with 2011 Census data to create a weighted average of the risk of automation across all jobs in each local government area (LGA).
The merged data was used to produce the graphs, which show the probability of job losses due to automation in each LGA in each of the eight ANZSCO ‘major groups’ .
The five most common jobs in each LGA are based on 2011 Census data; their risk of automation is based on CEDA’s research.
Note that the data takes no account of the new jobs automation will create in fields such as artificial intelligence, robotics, and machine learning.