Nighttime lights imagery is obtained using a low-light sensor on a satellite that captures quantitative measurements of light emissions at night. The value of this data is that artificial light emissions highlight human activity. It provides a unique perspective on the spread of our activities across the world. It has been used to estimate economic activity, as an alternative to traditional economic metrics such as Gross Domestic Product (GDP). Previous studies using nighttime lights data focused on the relationship between light intensity and GDP to determine if light intensity correlated strongly with GDP at national and subnational levels. In this study, annual composites of nighttime lights (for 2011, 2016, 2019 and 2021) and Landscan population estimates were used to create a model, using the Random Forest algorithm, that predicted Gross Value Added (GVA) at a one-kilometre spatial resolution for the City of Johannesburg – a metropolitan municipality located in Gauteng, South Africa that is the largest contributor to the country’s economy. The predicted model had a Spearman’s coefficient of 0.88 and an RMSD of 2468.54 Rands Million. The results in the predicted GVA showed a trend of increasing GVA from the period between 2011 and 2019, however, there was a noticeable decline in GVA and a contraction of the spatial spread of economic activity between 2019 and 2021. This coincides with the impact of Covid-19 and resulting lockdown measures, as well as the ongoing electricity interruptions, on the economy.
|