Census data used to produce the sample populations were preprocessed to remove nongeographical census regions and ensure consistency across different spatial scales. Australian census data are publicly available through the ABS’ online platform Census TableBuilder. To generate our sample populations, we used a subset of demographic data relating to the following characteristics of interest: dwelling location, employment status/work location, household composition, sex, and age. To gather this information, we accessed the following datasets: (i) statistical area level 1 (SA1) [usual residence (UR)] by AGEP (population count by age) and SEXP (population count by sex), (ii) SA1 (UR) by CDCF (count of dependent children in family) and NPRD (number of persons usually resident in dwelling), and (iii) SA1 (UR) by employment destination zone (DZN) [place of work (POW)].

All exported data are subject to perturbation to preserve the anonymity of individuals as per the Australia Census and Statistics Act 1905. The relative effect of perturbation becomes more prominent with the increased parsing of data. This effect is most severe in the TTW networks, as these break the population into the smallest groups.

The net effect of these perturbations can lead to inconsistencies between data amalgamated for different spatial partitions. For example, accumulating commuter data on the level of (small) SA1 partitions cannot reproduce the ABS-provided statistics over (larger) SA2 partitions and can lead to the creation of nonexistent edges in the SA2 (UR) to SA2 (POW) TTW network.

To avoid artificial links in the disease transmission network, we removed those SA1 to DZN entries that could not be accounted for on the level of SA2 [i.e., that produced nonexistent edges when amalgamated to the scale of SA2 (UR) to SA2 (POW)]. Because of a change in the ABS procedure for introducing perturbations into the 2016 data, some additional processing was required to ensure consistency between 2011 and 2016. This procedure involved the recovery of TTW network edges by sampling over several additional ABS datasets. We will release this modified dataset and outline our sampling procedure in detail in our forthcoming publication (41).

Note: The content above has been extracted from a research article, so it may not display correctly.

Please log in to submit your questions online.
Your question will be posted on the Bio-101 website. We will send your questions to the authors of this protocol and Bio-protocol community members who are experienced with this method. you will be informed using the email address associated with your Bio-protocol account.

We use cookies on this site to enhance your user experience. By using our website, you are agreeing to allow the storage of cookies on your computer.