Data scarcity to strategic insight: monitoring global SDG progress using big earth data and AI

Evidence-informed sustainable development goals (SDGs) remain constrained by incomplete, geographically uneven and temporally sparse reporting. This study implements a systematic approach using 20 peer-reviewed Big Earth Data and artificial intelligence (AI)-derived datasets to address monitoring gaps across seven SDGs. By integrating multi-source geospatial datasets, AI-based methods, and established validation procedures, we assess regional imbalance and temporal patterns of SDG progress at the global scale. Across the selected variables, conventional sources cover an average of 61.7% of countries or regions, whereas the spatial products provide near-global coverage at resolutions ranging from 10 m to 1°, closing an estimated 38.3-percentage-point geographic gap. They also enable trend assessment for all 20 variables, compared with 16 using the official comparison data. Under the baseline specification, 10 variables are improving, nine are deteriorating and one is stable. Nine of eleven variables associated with SDGs 2, 6, 7, and 11 improve, whereas seven of nine associated with SDGs 13–15 deteriorate. These findings demonstrate how high-resolution Big Earth Data and AI can complement conventional statistics, improve spatially explicit SDG assessment, and support more targeted, evidence-based policy responses.