Published in the Latin American Journal of Central Banking
This repository contains the code supporting the paper: "GDP Nowcasting: A Machine Learning and Remote Sensing Data-Based Approach for Bolivia".
This research presents an innovative nowcasting strategy for GDP in developing countries, addressing the common challenge of delayed official statistics. Focusing on Bolivia, where monthly economic growth indicators are published with up to a six-month lag, the proposed method successfully reduces this delay to just two months. The approach leverages machine learning techniques, combining traditional economic indicators with features extracted from satellite imagery.
Model robustness is thoroughly validated using multiple criteria, including performance benchmarks against conventional econometric models and sensitivity analysis across diverse feature sets. In addition to offering new insights into Bolivia’s economic dynamics, this framework provides a replicable methodology for countries facing similar data constraints.