National residential asbestos heatmap by predictive modelling

This approach to market seeks the provision of a national residential asbestos heatmap (the heatmap), developed using a variety of datasets with location based residential information. Several datasets will be provided by the Asbestos Safety and Eradication Agency (the Customer), including government information, commercial testing, and research data on the ACM status of residential properties. Additional datasets required for the heatmap will be provided as necessary by the Supplier, and as agreed with the Customer.

The heatmap will show the likelihood of asbestos containing materials (ACMs) being present or absent in the residential environment and is to be developed using artificial intelligence (AI) and data analytics approaches that incorporate the differing data formats of the available datasets. Ultimately, this should include an estimation of the number of homes in Australia that are predicted to still contain ACMs. The outputs (including the heatmap, associated products, project methodology and report) will need to be fit-for-purpose for our jurisdictional stakeholders as they work towards implementation of the National Strategic Plan for Asbestos Awareness and Management (NSP 2019–2023).

The Supplier will need to undertake an initial data clean-up process to transform all the datasets provided by the Customer into the appropriate format for their chosen approach. The Supplier is encouraged to consider what other datasets may be relevant to developing the heatmap and propose a rationale to the Customer for their inclusion in the heatmap. The datasets will only be considered for use if the Supplier can independently gain access to them and make them available to the Customer for (current or future) use.

The Supplier will deliver this project using an appropriate AI and analytics approach. It is expected that the Supplier will analyse datasets to understand, train and use fit-for-purpose AI models that are able to predict the presence or absence of ACMs in the residential environment. While the main input datasets contain asbestos information from expert opinion or laboratory analysis of ACMs from different locations within a home (e.g. outdoor fence, roof or cladding; indoor walls, ceilings or floors), the heatmap will need to integrate this as a whole-of-home prediction for ACMs.