4661.0 - Using electricity data to understand COVID-19 impacts , 2020  
ARCHIVED ISSUE Released at 11:30 AM (CANBERRA TIME) 02/10/2020  First Issue
   Page tools: Print Print Page Print all pages in this productPrint All

Interpreting Electricity Data

There are a number of challenges in using and interpreting electricity data generated by smart meters. Changes in business conditions and behaviour can be reflected differently in the data, and there is variety both between and within industries. As a result of lockdowns, some businesses may experience a decline in activity and a decline in electricity usage, while for other businesses, electricity usage may be unaffected. Additionally, for some businesses, a decline in electricity usage may coincide with a reduction in employment, while for others, employees may have shifted from working at the business location to working from home.

Similarly, a range of factors affect household electricity usage and patterns, and these vary from household to household. While clustering and other methods can be used to predict household behaviour, these methods are not perfectly accurate, and further data is required to enable quality assessment. While the data analysis suggests that a higher proportion of people were staying home from mid-March, this has a number of possible causes, such as people working or studying from home more, people losing their jobs or having their hours reduced, and people reducing their outside activity as governments introduced lockdown measures.

Electricity usage is affected by a number of factors, one of the most important of which is the weather (air conditioning in warm months, and heating in colder months). This is important to consider when interpreting trends, and the relationship between weather and electricity usage is an area for future investigation. Another important consideration is the difference between electricity purchasing and usage. The available data only includes electricity purchased from the grid, and does not include household and business consumption of self-generated electricity from solar panels.

The data presented in this report only cover inner Melbourne for several months in 2019 and 2020, and the results cannot be generalised Australia-wide. While Victoria has universal smart meter penetration, this is not the case for other States and Territories and this may present a methodological challenge if the ABS acquires data from beyond Victoria in the future.

Data quality issues also impact analysis, particularly in relation to business locations. When integrating smart meter data for business locations with the ABS Business Register, a match could be found for a proportion of these (67%). Additionally, some meters’ data matched with multiple businesses in different industries. Devising a method for apportioning the electricity usage presents a challenge.