Coming from TableBuilder?

If the table you need doesn't exist anywhere, TableBuilder can build it. You choose the variables, the categories and the areas: age by income by occupation for every SA2 in a state, say, or, in Pro, a custom area assembled from mesh blocks. Tables can run to tens of millions of cells, and Census data goes back to 2006.

It's also free. Census TableBuilder Basic comes with an ABS account, and the ABS lists Pro as free for members of a registered organisation. For researchers and analysts who know the Census dictionary, nothing else gives you this much control, and Censible doesn't try to replace it.

That control takes effort. You need an account, a working knowledge of Census variables, and patience with large tables, which can queue for minutes or hours. Census TableBuilder holds categories rather than raw values, so medians and averages have to be worked out afterwards in a spreadsheet.

Every cell carries a small random adjustment to protect privacy, so totals made by adding cells drift from the real total. And each census is its own dataset on its own boundaries, so comparing 2016 with 2021 for the same place means reconciling two sets of areas yourself. The ABS plans time series datasets for selected variables with the 2026 release, which will help.

Censible does that preparation once, on 618 metrics, most of them drawn from the ABS's community profile tables: population and age, income, housing, households, work, education and more. Medians are worked out for you, including for areas you draw, from counts added up first. The 2011, 2016 and 2021 censuses sit on the same 2026 boundaries, and the figures come back in a web app, as CSV, through a JSON API or to an AI assistant, with the source and method attached.

What you give up is the free choice of cross-tabulations. If you need a combination of variables that isn't among our metrics, TableBuilder is the right tool.