Solar machine learning

A good understanding of the solar PV network across the Sunshine Coast region has previously been limited to the solar generated electricity output per postcode. While the data captured provides some useful insights more detailed information was required to better inform community uptake trends.

With new technologies such as ‘machine learning’ (which uses the artificial intelligence (AI) of computer systems to analyse and draw links to patterns of data) has enabled Council to map solar PV systems across the Sunshine Coast. Machine learning is an emerging work process that can extract value from data such as aerial imagery by recognising data patterns and making predictions. Council regularly commissions the capture of aerial imagery and generates data from it for inclusion in the corporate mapping system.

The mapped solar panels information can improve decision-making when combined with other corporate mapping layers and databases such as buildings and facilities.

The application of machine learning technology has produced region-wide solar panel layers for nominated years, detected with an approximate accuracy of 90%. As we continually learn more about machine learning the level of accuracy should improve. To date, layers have been created for 2015, 2019 and 2020 with 2021 underway.

Solar panel trend analysis can be used to inform renewable energy planning and targeted strategies for enhanced, region-wide solar device uptake. Temporal comparisons will enable the measurement of on-ground outcomes both retrospectively and in response to solar promotion campaigns.