The National Strategic Plan for Asbestos Awareness and Management 2019-2023 (NSP 2019-2023) commits Commonwealth, state and territory governments to work together to eliminate asbestos-related diseases in Australia by preventing exposure to asbestos fibres. The Asbestos Safety and Eradication Agency’s (the Customer) key function is to co-ordinate the implementation of NSP2019-2023. NSP 2019-2023 has a suite of national targets one of which is to develop an evidence-based national picture that assesses the likelihood of asbestos containing materials (ACMs) being present in the residential environment.

The Customer aims to deliver a national residential asbestos ‘heatmap’ by predictive modelling which will indicate the concentration of residential properties containing asbestos within geographic regions in Australia (for example, by local government area). The evidence-based national picture will be used to inform policy development and initiatives such as identifying optional locations for future disposal facilities, disaster recovery planning and targeted home renovator information campaigns.

In order to complete this predictive modelling, we need to source various datasets with differing characteristics, including reliable information on the presence or absence of asbestos in the residential environment. Various datasets have already been sourced with differing characteristics, however more target data is required to build an accurate residential asbestos heatmap.

Datasets containing both target and features attributes are rarer in the residential sector, compared with the commercial or government sectors, where there is a legal obligation to maintain this type of information in asbestos registers and management plans. To date data has been sourced data from: jurisdictional housing associations, ACM Check app data, asbestos mapping and redevelopment information collected by local councils, Education Department asbestos registers, ACT Mr Fluffy pre-demolitions inspections, Australian Bureau of Statistics Housing Survey data, CoreLogic and the GeoScience Australia’s National Exposure Information System (NEXIS). Using the year of construction of homes (between 1950 and 1990) as a predictor, a basic heatmap has been developed showing suburbs with the highest likelihood of ACMs being present using the NEXIS dataset. However, a number of issues were raised with this approach for example, it will capture a house that was built during that period but has since had all its asbestos removed but will not capture a house built before that time and later renovated using asbestos. For these reasons, feature-types maps are not considered sufficiently reliable to enable governments to make policy decisions regarding resource allocation. Sophisticated modelling is required to make these maps more meaningful and the outcome fit-for-purpose.