data release report - pitneybowes.com · g-naf (geocoded national address file) is the...

Product:

Prepared: August 2017

G-NAF Data Release Report August 2017

Revision History Date Version Change Coordinator

August 2017 1.0 Initial Version Anthony Hesling

Disclaimer PSMA Australia believes this publication to be correct at the time of printing and does not accept responsibility for any consequences arising from the use of information herein. Readers should rely on their own skill and judgement to apply information to particular issues. This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced by any process without prior written permission of PSMA Australia Limited.

G-NAF - Data Release Report August 2017

Page i

Contents Introduction ............................................................................................................................................ 1

Document Purpose .............................................................................................................................. 1

Audience .............................................................................................................................................. 1

Support ................................................................................................................................................ 1

Overview ................................................................................................................................................. 2

Updates and Issues ............................................................................................................................... 3

Updates ............................................................................................................................................... 3

Issues ................................................................................................................................................ 14

Summary Statistics ............................................................................................................................. 15

Future Considerations ........................................................................................................................ 17

Annex A – Detailed Statistics by Jurisdiction .................................................................................. 18


Page ii

Introduction G-NAF (Geocoded National Address File) is the authoritative physical address index for Australia. It contains the state, suburb, street, number and coordinate reference or geocode for street addresses in Australia. G-NAF uses existing and recognised address sources (referred to as contributors) including the state and territory government land records, as well as address data from the commonwealth government and other national sources. For more information please refer to the G-NAF Product Description.

Document Purpose This document provides a description of the key issues encountered and production changes applied during the assembly of this update. In addition, overarching release statistics and general change volumes are also included. This document should be read in conjunction with the accompanying G-NAF Product Description.

Audience This document is made available to PSMA staff, data contributors, direct clients and end users. It is intended to be used by business managers, data managers and analysts as a guide and reference to understanding the content and variations within a data release.

Support PSMA Australia Support PSMA Australia Limited www.psma.com.au/contact


Page 1

http://www.psma.com.au/contact

Overview • Nationally, this update of G-NAF shows an increase of 53,441 addresses overall which

equates to an increase of 0.38%. The total number of addresses in G-NAF now stands at 14,264,626 of which 13,648,135 or 95.68% are principal.

• Confidence levels nationally show a slight improvement in confidence level 2 addresses which have increased by 36,255 and now represent 65.80% of principal addresses. Confidence 1 addresses have increased slightly since the May 2017 release from 14.80% to 14.83% and confidence 0 addresses have decreased from 19.68% to 19.37% of addresses.

• There was an increase of 5,433 address level, a drop of 4,689 in street locality addresses and a drop of 4,161 in locality level addresses. The reliability of principal addresses in G-NAF is now made up of 96.24% address level, 3.54% street locality addresses and 0.22% locality addresses.

Updates

• A number of updates and quality improvements have been undertaken for the August 2017 release of G-NAF, these include:

o uniform geocode placement o allocation of postcodes o Queensland implementation of jurisdictional alias addresses o geocode level improvement by resolving locality only addresses o changes to Localities names and boundaries

• There has also been a series of minor updates to processing of addresses where minor anomalies have been identified and in each case affects a small number of addresses in a localised area.

• For commercial clients who use the LYNX Proprietary Binary Format, a new version of the LYNX Data Importer software is available for download (v1.5). The August 2017 release of G-NAF will require this new version of the LYNX Data Importer.

Issues

• There have been no significant production issues in the August 2017 release of PSMA G-NAF data.


Page 2

Updates and Issues Updates Uniform Geocode Placement

Overview As outlined in the May 2017 release notes, the August 2017 release includes the implementation of PROPERTY CENTROID (PC) geocode types in ACT and Victoria as an alternate placement for the existing FRONT CENTRE SETBACK (FCS) geocode types. As a result, the majority of address level addresses in ACT and Victoria now have TWO address level geocodes allocated to them. This is not dissimilar to the situation in South Australia where many rural addresses have a PROPERTY CENTROID (PC) geocode type and a PROPERTY ACCESS POINT (PAP) geocode type. The addition of PROPERTY CENTROID (PC) geocode types in ACT and Victoria has been undertaken to support those users seeking a more uniform approach to address level geocodes in G-NAF. Most importantly, there is no change for those users who wish to continue with the geocode placement used prior to this update. Note: The default geocode table will deliver the same outcomes in terms of geocode placement in the August 2017 release as it did in May 2017.

Background The G-NAF data model accommodates multiple address level geocode types for a single address (eg. access point, property/parcel centroid etc.). During the G-NAF production process, address level geocodes are generally based on the geocodes provided by state and territory jurisdictions. As a result, the differences across common address geocode types reflects the respective allocation practices in place in each jurisdiction. In practice this means that currently G-NAF does not contain a single uniform address level geocode type across all states and territories. As part of the collaboration activities with the Australian Government, PSMA undertook to add additional geocodes into G-NAF to deliver the option for users to utilise a consistent address level geocode type across all states and territories. PSMA had identified that most jurisdictions capture a PROPERTY CENTROID (PC) geocode type, with approximately 70% of address level geocodes across Australia being of this type. Geocodes in the ACT and Victoria being the key exceptions using FRONTAGE CENTRE SETBACK (FCS). The implementation of PROPERTY CENTROID (PC) geocode types for ACT and Victoria DO NOT apply to the following types of addresses:

• CONFIDENCE = -1 (retired address);

• GAP GEOCODED (GG) geocode type; or

• PROPERTY CENTROID MANUAL (PCM) geocode type.

ACT In the ACT, those addresses with a FRONT CENTRE SETBACK (FCS) geocode type not associated with strata development have now been allocated a property centroid geocode in addition to their existing geocode. (ie. the address now has two records in the ADDRESS_SITE_GEOCODE table). The number of principal addresses affected by this change is 191,395. For addresses associated with strata developments, the classification of the geocode has been simply changed from FRONT CENTRE SETBACK (FCS) to BUILDING CENTROID (BC) as this more accurately reflects the nature of the geocode in that it has generally been placed within the building


Page 3

footprint or in very close proximity to the building. The number of principal addresses affected by this change is 18,584. In a small number of cases, the PROPERTY CENTROID (PC) geocode associated with the FRONT CENTRE SETBACK (FCS) geocode has not been generated for a reason yet to be resolved. These includes the primary addresses associated with strata developments. This issue will be resolved in the next release. The following examples are indicative of the changes that have been made in the ACT.

Example 1

Example 2


Page 4

Example 3

Example 4


Page 5

Victoria In Victoria, 99% of addresses with a FRONT CENTRE SETBACK (FCS) and PROPERTY ACCESS POINT SETBACK (PAPS) geocodes have been allocated a PROPERTY CENTROID (PC). Like in the ACT, there are a very small number (ie. less than 1%) where the PROPERTY CENTROID has not been created and are subject to further investigation. Unlike the ACT, however, no exceptions have been made for strata developments given many of the geocodes for these developments had been placed in a front centre setback position. The number of addresses where the additional PROPERTY CENTROID geocode types have been created totals 3.3 million. The following examples are indicative of the changes that have been made in Victoria. Several examples of the outstanding issues are also shown in examples 5, 6 and 7.

Example 1


Page 6

Example 2

Example 3


Page 7

Example 4

Example 5


Page 8

Example 6

Allocation of Postcodes

The process of allocating a postcode to every address has been updated for the August 2017 release. The process has been made more consistent in the allocation, with all addresses in a gazetted suburb/locality given the same postcode. The inclusion of a postcode in G-NAF is primarily based on the most commonly used postcode for all the addresses within in each gazetted suburb/locality. One of the exceptions is the addresses located in the area allocated to postcode 3004 for St Kilda Road, where the majority of addresses would be in the suburb of MELBOURNE (3000). Postcodes are used by Australia Post to facilitate the efficient processing and delivery of mail. There is no mandatory requirement for the inclusion of a postcode in an address (AS4590). The overall impact of changes to address postcodes has been minimal with a total of 23,126 addresses with a changed postcode between the May 2017 and August 2017 releases.

ACT 106

NSW 8,430

NT 239

OT 0

QLD 6,183

SA 2,320

TAS 333

VIC 4,610

WA 905

TOTAL 23,126

Count of addresses affected by a postcode change


Page 9

The examples below show the general nature of the addresses that have changed their postcode in this release as a result of the revised process. In most cases, the addresses are near locality boundaries. In the image to the left, the changed postcodes are along a road. In the image to the right there are in excess of 1000 addresses on one geocode which results 25% of the total addresses in Victoria where the postcodes have changed. In the centre image most of the addresses are near the locality boundaries, however there is one address in the centre which certainly highlights the benefits of this new approach in terms of identifying the most appropriate postcodes.

Queensland implementation of jurisdictional alias addresses

In the August 2017 release, the GNAF process for Queensland now takes into account the alias-principal status allocated to some addresses supplied by the jurisdiction. As such, an additional 51,000 addresses are now alias addresses in Queensland and therefore, further reducing the number of “duplicated addresses” at a site. These alias addresses exist primarily where a property is adjacent to more than one street (eg. corner property). Examples of situations where this has occurred for provided in the following images.

Example 1


Page 10

Example 2

Example 3

Geocode level improvement by resolving Locality-only addresses

This update has seen the single largest reduction in locality-only addresses (ie. addresses with only a locality geocode) for the past 10 updates of G-NAF. There was a reduction of some 4000 locality-only addresses (ie. 12%), with most of this occurring in Queensland where the reduction was 28%. This has involved the creation of a large number of rules to either:

• correct the locality name and/or the street names of addresses that do not align with the gazetted locality and road names, or

• create primary-secondary relationships where the addresses are associated with internal road names with private developments such as retirement villages.

The reduction in locality-only addresses was significantly assisted by the inclusion of information provided by Commonwealth government agencies. Corrections to localities and road names are reflected in the street-locality alias tables.


Page 11

Several examples of the type of changes made are provided below.

Example 1

Example 2


Page 12

Example 3

Changes to Localities names and boundaries

The pre-release notes for Administrative Boundaries identified changes to localities in New South Wales, Queensland, Victoria and Western Australia for the August 2017 release. The changes in Victoria include some significant boundary changes in an area to the west of Melbourne. These changes included the addition of 11 new localities. As a result, a number of addresses have changed their locality name and as a consequence their ADDRESS_DETAIL_PID. The image below provides an overview of the boundary changes made.


Page 13

LYNX Data Importer software update (commercial clients only)

The August 2017 release of G-NAF will require the new version (1.5) of the LYNX Data Importer software to use the LYNX Proprietary Binary Format. The updated version is compatible for the current and all previous releases of G-NAF. The LYNX Data Importer software is available to commercial clients for download from PSMA’s Online Data Delivery system. Once logged in, a link to the installers is available on the Obtain PSMA Data > G-NAF pages. Please see the Client User Guide for more information on obtaining and installing the software. If you have any questions or issues with the update to the LYNX Data Importer please contact PSMA Support ([email protected] ).

Issues There are no significant production issues. However, PSMA has identified that approximately 200,000 addresses have a changed LEGAL_PARCEL_ID. Whilst the majority of the changes corrected the previous value, further investigations are required to continue the improvement of LEGAL_PARCEL_ID allocation.


Page 14

https://services.psmasystems.com.au/lynx/smarty/application/sdw_online/login.php

https://services.psmasystems.com.au/lynx/smarty/application/sdw_online/login.php

mailto:[email protected]

Summary Statistics Note: The statistics below do not include “retired” addresses (ie. Confidence =-1).

Geocode Reliability of Principal Addresses Only

Reliability No of Addresses Percentage Change Since

Feb 2017

1 (GPS derived address level) 0 0.00% 0

2 (Within address site boundary) 12,981,058 95.11% 5,525

3 (gap geocoded) 153,851 1.13% -92

4 (street level) 482,684 3.54% -4,689

5 (locality level) 30,542 0.22% -4,074

6 (topo level) 0 0.00% -87

No geocode 0 0.00% 0

Total 13,648,135 100.00% -3,417

Address Confidence Levels of Principal Addresses Only

Confidence level Count Percentage of Total

Change Since Feb 2017

Addresses with a Confidence of 0 2,643,426 19.37% -42,968

Addresses with a Confidence of 1 2,023,624 14.83% 3,296


Total 13,648,135 100.00% -3,417

Geocode Reliability of Addresses (Principals and Aliases)

Reliability No of Addresses Percentage Change Since

Feb 2017

1 (GPS derived address level) 0 0.00% 0

2 (Within address site boundary) 13,593,055 95.29% 62,710

3 (gap geocoded) 155,159 1.09% -108

4 (street level) 485,820 3.41% -4,987

5 (locality level) 30,592 0.21% -4,087

6 (topo level) 0 0.00% -87

No geocode 0 0.00% 0

Total 14,264,626 100.00% 53,441


Page 15

Address Confidence Levels of Addresses (Principals and Aliases)

Confidence level Count Percentage of Total

Change Since Feb 2017

Addresses with a Confidence of 0 3,070,634 21.53% -299



Total 14,264,626 100.00% 53,441


Page 16

Future Considerations Address-Property Linkage

Since its initial production in 2004, the G-NAF data model has included a provision for a linkage between properties and addresses. This linkage has been maintained in an ad-hoc manner leading to varying levels of quality in the data in the respective linkage fields. Following on from the successful upgrade to the production processes to improve address and cadastral parcel linkage data, PSMA is seeking to develop similar production processes to deliver high quality address and property linkages. The commencement of this development work will be subject to wider stakeholder approval and will likely be implemented for the November 2017 production cycle. It is not expected that the G-NAF data model will vary as a result of this production process development.


Page 17

Annex A – Detailed Statistics by Jurisdiction Victoria

Victorian addresses represent 25.14% of the national total. Victoria has had an increase of 18,270 addresses for this release. Both address reliability and confidence remain largely unchanged in Victoria for the August release.

Confidence - Vic No of Addresses % Change from

Previous Release

% Change

Addresses with a Confidence of 0 543,328 15.84% 3,267 0.01%


Addresses with a Confidence of 2 2,362,983 68.87% 12,352 -0.01%

Level Geocoded - Vic No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 3,372,542 98.30% 18,173 0.01% Addresses geocoded to Street level 54,528 1.59% -10 -0.01% Addresses geocoded to Locality level 3,760 0.11% 107 0.00%

New South Wales

New South Wales addresses represent 30.92% of the national total. There was an increase of 16,183 principal addresses for NSW. Both address reliability and confidence show improvement for the August release in NSW.

Confidence - NSW No of Addresses % Change from

Previous Release

% Change

Addresses with a Confidence of 0 820,852 19.45% -620 -0.09%


Addresses with a Confidence of 2 2,729,730 64.68% 12,290 0.04%

Level Geocoded - NSW No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 4,021,370 95.29% 19,995 0.11% Addresses geocoded to Street level 186,327 4.42% -2,995 -0.09% Addresses geocoded to Locality level 12,522 0.30% -817 -0.02%


Page 18

South Australia

South Australian addresses represent 7.79% of the national total. South Australia has had an increase of 1,130 principal addresses. Both address confidence and address reliability shows a slight improvement for South Australia in the August release.

Confidence - SA No of Addresses % Change from

Previous Release

% Change

Addresses with a Confidence of 0 251,064 23.63% -2,588 -0.27%



Level Geocoded - SA No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 1,018,920 95.89% 2,265 0.11% Addresses geocoded to Street level 40,905 3.85% -732 -0.07% Addresses geocoded to Locality level 2,784 0.26% -403 -0.04%

Tasmania

Tasmanian addresses represent 2.39% of the national total. There was a small increase of 422 principal addresses for Tasmania which continues to show improving trends for both reliability and address confidence for the August 2017 release.

Confidence - Tas No of Addresses % Change from

Previous Release

% Change

Addresses with a Confidence of 0 58,821 18.01% 4 -0.02%


Addresses with a Confidence of 2 218,819 67.02% 441 0.05%

Level Geocoded - Tas No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 307,540 94.19% 662 0.08% Addresses geocoded to Street level 18,132 5.55% -153 -0.05% Addresses geocoded to Locality level 845 0.26% -87 -0.03%


Page 19

Western Australia

Western Australian addresses represent 10.44% of the national total. Western Australia has had an increase of some 4,560 principal addresses. Both confidence and address reliability show a slight improvement for the August 2017 release.

Confidence - WA No of Addresses % Change from

Previous Release

% Change



Addresses with a Confidence of 2 1,003,368 70.41% 5,645 0.17%

Level Geocoded - WA No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 1,375,013 96.49% 5,635 0.09% Addresses geocoded to Street level 46,327 3.25% -540 -0.05% Addresses geocoded to Locality level 3,720 0.26% -535 -0.04%

ACT

ACT addresses represent 1.57% of the national total. There was an increase of 981 principal addresses in ACT. Address confidence shows a slight improvement in ACT while the address reliability has fallen slightly for the August 2017 release.

Confidence - ACT No of Addresses % Change from

Previous Release

% Change




Level Geocoded - ACT No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 212,723 99.02% 939 -0.02% Addresses geocoded to Street level 2,049 0.95% -6 -0.01% Addresses geocoded to Locality level 48 0.02% 48 0.02%


Page 20

Queensland

Queensland addresses represent 20.95% of the national total. There was a reduction of some 45,409 principal addresses in Queensland as a result of identifying and implementing alias addresses. As a result of the additional alias identification, the address confidence match rates show a good improvement for Queensland for the August 2017 release while address reliability continues to show steady progress.

Confidence - QLD No of Addresses % Change from

Previous Release

% Change


Addresses with a Confidence of 1 422,379 14.77% -4,436 0.08%

Addresses with a Confidence of 2 1,771,785 61.97% 191 0.98%

Level Geocoded - QLD No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 2,724,598 95.30% -42,668 0.02% Addresses geocoded to Street level 128,003 4.48% -251 0.06% Addresses geocoded to Locality level 6,370 0.22% -2,490 -0.08%

Northern Territory

Northern Territory addresses represent 0.78% of the national total. Northern Territory had a small increase of 446 addresses. Address confidence and address reliability show a slight improvement for the August 2017 release.

Confidence - NT No of Addresses % Change from

Previous Release

% Change




Level Geocoded - NT No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 99,535 93.84% 432 0.01% Addresses geocoded to Street level 6,125 5.77% -2 -0.03% Addresses geocoded to Locality level 406 0.38% 16 0.01%


Page 21

Other Territories

Other territories addresses represent 0.02% of the national total. Other Territories remains unchanged since the May 2017 release.

Confidence - OT No of Addresses % Change from

Previous Release

% Change



Addresses with a Confidence of 2 22 0.72% 3 0.09%

Level Geocoded - OT No of Addresses % Change from

Previous Release

% Change

Addresses geocoded to Property level 2,668 87.68% 61 0.86% Addresses geocoded to Street level 288 9.46% 5 0.04% Addresses geocoded to Locality level 87 2.86% -26 -0.90%


Page 22

data release report - pitneybowes.com · g-naf (geocoded national address file) is the...

Documents