Abstract

Stats NZ is looking at different ways to form the basis for future statistical production, which can produce existing statistics in a more efficient way, while also being flexible to future requirements and cross-domain insights using integrated social, economic and environment registers and data.  Stats NZ is transforming it’s siloed statistical data infrastructure to be more joined up with the development of an Integrated Statistical Data System (ISDS). 

The ISDS is based on the internationally recognised Register-based Statistical System (RSS). The basic concept of a statistical system comprised of three central domains (businesses, people, and places) is the same. However, New Zealand does not have administrative registers that can underpin the required statistical infrastructure. Instead, each domain has a central statistical spine, or system of spines, developed from independent, but integrated, data sources. These central statistical spines represent, for statistical purposes, the role of statistical registers in a Register-based Statistical System.

Increasing the use of administrative data is critical for delivering future census type information and meeting the data needs of Aotearoa New Zealand. A Census counts people once, only once and in the right place.

For Stats NZ much has changed since the introduction of the Statistical Location Register (SLR) in 2016. Our understanding of, and the availability of, property data has increased significantly. The absence of key property entities in the SLR means change that impacts a dwelling is either not known or cannot be placed in an unambiguous and clear context. Currently this gap is worked around by recreating a dwellings dataset every five years using census manual update processes.

Stats NZ is moving to a more property-centric Places Index which introduces the key entities of property and building and links these entities with address and dwelling in a more precise and accurate manner, consistent with the definitions of the data. This enables changes identified in admin data to be matched to the exact entity affected, allowing the impact on an exact dwelling to be determined.

These new entities also allow the introduction of property scenarios. These scenarios model the real-world types of property that exist, which largely determines the type of dwelling that a property can have. Scenarios also allow the record-level relationships for individual properties to be unpicked, further driving out ambiguity and simplifying the connections.

The Places Index spans both commercial and residential properties, allowing connections to be made to both the Statistical Business Register (SBR) and the Administrative Population Model (APM). Initially the Places Index will focus on delivering and maintaining an admin data dwelling register to support future census requirements.

This presentation explores Stats NZ’s Places Index development, demonstrating a more property-centric model linking address to property to building to dwelling and how these changes enable admin data to identify dwellings to ensure a unique set of dwellings can be extracted at any point in time and to support the production of various dwelling statistics for Aotearoa New Zealand.