This dataset's primary purpose is to reproduce the results in Section 4.4 in A sparse linear algebra algorithm for fast computation of prediction variances with Gaussian Markov random fields by Zammit-Mangion and Rougier, which can be found here. All data here was produced by the Australian Bureau of Statistics and is freely available under the Creative Commons Attribution 2.5 Australia License. The ZIP file contains two folders, census_data and shapefiles. These are discussed in turn. census_data This folder contains 9 CSV files. * The filenames containing "SA1" contain data at the SA1 level, while the filenames containing "SA2" contain data at the SA2 level. * The filenames containing "B02" contain selected median and averages of income (not used in the study). * The filenames containing "B16" contain data on education. Note that "B16A" and "B16B" refer to "Book A" and "Book B" of the same dataset. * The filenames containing "B26" contain family income data. The folders "About DataPacks" and "MetaData" contain more information about the datasets. To access entire DataPacks, please visit the DataPack webste. The other file in this folder is censuscounts_mb_2011_aust.csv. This contains the Census counts at the Mesh Block level and this data can be downloaded from here. shapefiles This folder contains shapefiles of the Mesh Blocks (MB) in New South Wales (NSW), those for the SA1 regions and those for the SA2 regions for all of Australia. These shapefiles are freely available from (here)[http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/1270.0.55.001July%202011].