Summary
Cambodia is not alone in facing capacity limitations in the production and timely release of key official statistics needed for data-driven policy decisions. This paper demonstrates that combining satellite-derived indicators (e.g., nighttime lights, NO₂ emissions, vegetation indices) with traditional high-frequency indicators in a machine learning framework significantly improves the accuracy of GDP nowcasts. Moreover, satellite data enables closer examination of subnational patterns, providing granular, near-real-time insights into economic activity. These findings highlight the potential of non-traditional approaches to complement conventional methods and strengthen macroeconomic surveillance in data-scarce environments.
Subject: Agricultural sector, Economic forecasting, Economic sectors, Health
Keywords: Agricultural sector, big data, machine learning, non-traditional data, nowcast, nowcasting, random forest, satellite, satellite data