An Introduction to Statistical Geography

Introduction

Most of the data contained within AURIN is spatially aggregated. This means that, rather than representing a single point in space, the data are either summed, or averaged, across an area.

There are some complex, standardised systems of aggregation within Australia. The Australian Bureau of Statistics (ABS) has a particularly complex system of aggregation, made more-so by the fact that they are in the process of transitioning to a new geographic system. This system underpins a lot of the statistical geography and reporting undertaken by other non-ABS organisations throughout Australia, to ensure maximal comparability.

This page provides a brief introduction to the statistical geographic standards within Australia

Pre 2011 ABS Geography

Prior to the 2011 Census, the ABS maintained a system of geographic aggregation called Australian Standard Geographic Classification

The principles of the main structure of the ASGC are that they are:

  • Mutually Exclusive – that at a specific level of aggregation, a point in space is only covered by one unit area at that level;
  • Mutually Exhaustive – that the entire area of Australia was covered by unit areas at every level (there are some exceptions discussed below); and
  • Perfectly Nested – that moving upwards through the system to higher levels of geography does not involve splitting of areas, that is, that a subunit fits entirely within a single higher level of aggregation.

The concept of nested hierarchies is shown below

Hierarchy

[Click to Enlarge]

The official base level unit, at which most data for the 2006 Census is published, underlying all of the hierarchies and levels of aggegration in the 2006 ASGC, is the Census Collection District (CCD) (although each of the CCDs themselves comprise Meshblocks (MBs), at which very little census data are published, other than total population counts). There were just over 38,000 CCDs in Australia for the 2006 Census, each comprising an average of about 225 households. Obviously as population density increases or decreases, the size of the CCD decreases or increases accordingly.

The CCD is the base level unit for a number of higher aggregations, and hierarchies of aggregations. These are shown in Figure 2. Boxes with complete lines represent a level of aggregation that covers all of Australia (i.e. Mutually Exhaustive), while dashed lines represent areas that only cover part of Australia. Boxes with shadows indicate a level of aggregation for which data is available in the AURIN portal.

2006ABS

[Click to Enlarge]

The Main ASGC Structure

The main ASGC structure is shown in Figure 1 (bounded in green), together with other aggregation hierarchies that the ABS supported during this period. The main structure was:

CCDs → Statistical Local Areas (SLAs) → Statistical Sub-Divisions (SSDs) → Statistical Divisions → (SDs) → States and Territories → (STEs) → Australia

In this structure, SLAs aggregate to form SSDs, and this aggregation principle continues up the remaining hierarchical levels.

At each hierarchical level, the component spatial units (e.g. SLAs) collectively cover all of Australia (i.e mutually exhaustive) without gaps or overlaps (i.e. mutually exclusive).

Statistical Region

The Statistical Region structure is shown in Figure 2 (bounded in blue). This structure was:

CCDs → SLAs → Statistical Region Sectors (SRSs) → Statistical Regions (SRs) → Major Statistical Regions (MSRs) → STEs → Australia

Local Government Areas

SLAs aggregate up to Local Government Areas (LGAs), which represent the boundaries between local/municipal governments across Australia (orange component of Figure 1). These aggregate up to State/Territory Level. There are a lot of datasets published at the LGA level within the AURIN portal, and this represents a very useful level of aggregation for analysis and mapping.

State Suburb Codes

CCDs aggegrate up to State Suburb Codes (SSCs), which represent gazetted suburbs or rural localities around Australia (orange in Figure 1)

Electoral Boundaries

CCDs also aggregate to both Commonwealth Electoral Divisions (CEDs) and State Electoral Divisions (SEDs), shown in purple in Figure 1. These levels of geographical division represent the lower house electoral boundaries for each state/territory, for federal elections and state elections respectively.

Postcodes

CCDs aggregate to Postcode level across Australia, which themselves aggregate seamlessly to State/Territory level (lilac box in Figure 1). Currently there are no datasets published at Postcode level within the AURIN portal

Urban Centre/Localities

Urban Centre/Localities (UCLs) are a non mutually exhaustive aggregations, which only cover part of Australia. The UCL Structure represents aggregations of CCDs to form defined areas according to specific population and land use criteria. Population counts (place of enumeration) from the latest Census of Population and Housing are used to define the UCL Structure, which means this classification structure is only defined at times of each census. The resulting areas are known as Urban Centres or Localities.

In broad terms, an Urban Centre is a population cluster of 1,000 or more people while a Locality is a population cluster of between 200 and 999 people. For statistical purposes, people living in Urban Centres are classified as urban while those in Localities are classified as rural.

Each Urban Centre/Locality has a clearly defined boundary and comprises one or more whole CCDs. Urban Centres/Localities are redefined at each Population Census.

2011 ABS Geography

Leading up to and following the 2011 Census, the ABS embarked on a rationalisation of the standard geographic boundaries and systems of aggregation used to report statistics. In addition to streamlining the process, the new Australian Standard Geographic System (ASGS) also ensured that all systems of aggregation were mutually exhaustive – previous systems where only part of Australia was covered (such as the UC/L) were redesigned so that all areas were covered by a categorisation at each level.

The base unit of the 2011 ASGS is the Statistical Area Level 1 (SA1), analogous to the collection district, although slightly smaller, housing between 200 to 800 persons, with an average of 400 persons.

The Main ASGS Structure

The main structure of the 2011 ASGS is represent in Figure 3 bounded in green.

2011ABS

[Click to enlarge]

This structure is:

SA1 → Statistical Area Level 2 (SA2) → Statistical Area Level 3 (SA3) → Statistical Area Level 4 (SA4) → States and Territories (STEs) → Australia

SA2s are a general-purpose medium-sized area built from whole SA1s. Their aim is to represent a community that interacts together socially and economically. In addition to aggregating to SA3, SA2s also aggregate to Significant Urban Areas (SUAs – shown in tan in Figure 3). They have a general population of 3000-25000, with an average of 10000 persons.

Where possible, SA2s have been designed around whole gazetted suburbs or rural localities.

SA3s provide a standardised regional breakup of Australia. SA3s are built from whole SA2s and do not cross STE boundaries. SA3s are designed to have populations between 30,000 and 130,000 persons. SA3s are often the functional areas of regional towns and cities with a population in excess of 20,000 or clusters of related suburbs around urban commercial and transport hubs within the major urban areas.

SA4 regions are the largest sub-State regions in the Main Structure of the ASGS. A minimum of 100,000 persons was set for the SA4s, although there are some exceptions to this. In regional areas, SA4s tend to have populations closer to the minimum (100,000 – 300,000). In metropolitan areas, the SA4s tend to have larger populations (300,000 – 500,000).

Labour markets were a key consideration in the design of SA4s. The reason for this is that Labour force data has two geographic components to it – the labour supply (where people live) and demand (where people work). For statistical purposes, it is ideal to maximise the extent to which the data output region spatially contains both sets of geographic locations.

Local Government Area

As with the 2006 ASGC, the ASGS maintains a level of LGA aggregation, in order to allow the reporting of statistics at Local Government and Municipal level (and allow inter-censal comparisons) shown in orange in Figure 3. These are constructed from SA2s.

State Suburb Codes

SA1s aggegrate up to State Suburb Codes (SSCs), which represent gazetted suburbs or rural localities around Australia (orange in Figure 3). These are roughly equivalent with SA2s, but are maintained to allow intercensal comparability.

Electoral Boundaries

SA1s also aggregate to both CEDs and SEDs, shown in purple in Figure 3.

Postal Areas

Postal Areas are ABS spatial units aggregated from SA1s to approximate postcodes (common postcodes are not a statistically rigorous geography). SA1s aggregate to Postal Areas across Australia, which themselves aggregate seamlessly to State/Territory level (lilac box in Figure 3)

Urban Centre/Localities

The UC/L system comprises a mutually exclusive/mutually exhaustive system within the 2011 ASGS. UC/Ls are aggregated from SA1s. The levels of aggregation are:

SA1 → UCL → Section of State Ranges (SSRs) → Section of State (SOS)