Policymakers, economists and senior executives are heavily reliant on economic data to make informed decisions. For instance, to measure economic activity, Gross Domestic Product (GDP) is the most common metric being used. However, In India these estimates are released approximately 7-8 weeks after the end of a quarter. Even other emerging market economies face serious problems of data lags, gaps and revisions which hamstring optimal policy decisions. This makes - a real-time assessment - of the state of the economy, a necessity.

With major advancements in high frequency data streams such as satellite imagery and anonymized smartphone location data, it has become possible to forecast what is right around the corner in newer, robust ways. These advancements have made it possible to move beyond traditional Forecasting that is concerned with predicting future trends and events, to Nowcasting, a real time analysis of our immediate present.

Through this article series, we explore applications of Nowcasting using high frequency and high-resolution satellite imagery, focussing on GDP construction. Imagine being an infrastructure player who needs to regularly make high-stake decisions around how the construction or built-environment is progressing across the country - are you willing to wait 7-8 weeks before you know what happened in the past?
Now you can leverage high frequency and high-resolution satellite imagery data to instantly know what happened just a few days ago and not wait for weeks on end. With availability of both radar and optical high-resolution EO satellites with a revisit rate of less than a week (sometimes, few hours!), we can get very precise estimates of how the built-environment has changed over a short period of time.

Using an automated change-detection algorithmic pipeline, we detect and isolate pixel level changes over time and draw out temporal trends of how construction unfolds in micro-pockets.

Among the extracted output, the red regions in Figures 1 & 2 are the processed output showing maximum change between the starting and end of the period of analysis. These outputs can be further processed to extract more meaningful metrics to power several key decisions around the built-environment that can’t and shouldn’t wait forever, anymore!