The agency has defended the study because it did not track people in real time, with a spokesman saying "the ABS has not tracked or conducted surveillance on any Australian".

The study used mobile phone data from two weeks in April 2016, with suburbs across the ACT and surrounding areas in NSW targeted to show the ways populations move throughout the day and at different times in the week.

First reported by freelance journalist Asher Wolf, the study has been made public through presentations given by demographer Andrew Howe at conferences in Adelaide and Redcliffe, Queensland last year.

Slides for the presentations claim that mobile device user data that was aggregated and anonymised is the solution to a data gap in showing temporary population estimates.

The ABS presentation said it had a "trial engagement with a telco company," but did not specify which company this was. Telstra has since confirmed that it was the company involved.

Privacy advocates have slammed the study and how the ABS had dealt with the data. Chair of Digital Rights Watch Tim Norton said the study was "irresponsible" by the ABS in light of controversy over the 2016 census.

"They're under a dark cloud of lacking consent in the way they engage with people's personal data so to do this study and then try and hide it, which is what it looks like, I think is incredibly irresponsible," he said.

Chair of the Australian Privacy Foundation David Vaile has compared the ABS' attitude around people's data to that of Facebook.

"It's a very contemptuous and patronising but also Facebook-like attitude to their right to keep things quiet or to not reveal the full extent of what they're doing," he said.

The ABS spokesman said individuals could not be identified from the data.

"The ABS received aggregate level telecommunications data from a telecommunications company to inform estimates of temporary populations. No individual information was provided to the ABS. Only aggregated counts of mobile telephone transactions were supplied in broad geographical areas," the ABS spokesman said.

The ABS said the information was obtained under the authority of the Census and Statistics Act, and that it is also bound by the Privacy Act. It said that it was not required to undertake a privacy impact assessment, because no personal data was used.

"Under this legislation, the ABS is bound to maintain the secrecy of information collected and protect people's confidentiality," the spokesman said.

The study was part of "ongoing efforts to use existing available data, rather than obtaining it directly from the population".

"This is accepted, common practice in many overseas countries."

The study was undertaken in the months leading up to the 2016 census, where the ABS was criticised for its plan to retain respondents' names and addresses for four years instead of the 18 months used in previous years. The period in the study was before mandatory data retention laws came into effect, but the ABS is one of the government agencies that can request metadata from telecommunications companies.

Telstra has also defended its role in the study.

"Telstra is firmly committed to the protection and security of our customers’ personal information, which is outlined publicly in Telstra’s Privacy Statement, and we comply with all relevant privacy legislation," a spokesman said.

"We do not share any information which identifies, or could reasonably identify, a customer with third parties unless we have the appropriate informed consent or are required to by law."

"For the ABS project Telstra supplied anonymised and aggregated data insights under the authority of the Census and Statistics Act 1905. No individual customer information was provided," the spokesman said.

The data provided by the company allowed the ABS to make hourly population estimates, showing how the population grew in the ACT during business hours, and dropped in surrounding areas of NSW at the same time.

It also showed that the number of people in the Belconnen Town Centre rose to around 10,000 each weekday, almost double the area's estimated resident population.

The data projected hourly population estimates in Bruce at the time of a Raiders v Tigers NRL match, including a breakdown of females and males. It also showed the number of people in the area at the time of the Groovin the Moo music festival.

The data was also able to project how many people attended AFL matches at the MCG on particular dates, as well as the population in beach towns on the NSW south coast over a long weekend.

"The pilot was conducted in the Australian Capital Region (the ACT and surrounds). Many of the people who use the services and infrastructure in Canberra do not reside in the ACT, yet Commonwealth funding is allocated to states and territories based on its usual resident population, not its service population," the ABS spokesman told Fairfax Media. This information could be used to make population-based funding decisions in areas with fly-in, fly-out populations, the spokesman said.