Economic policy questions increasingly depend on novel measurement: satellite imagery to track deforestation, digitized historical maps to study long-run development, and neural networks to predict poverty. This course surveys how economists build new spatial data from computational methods and deploy it for credible policy evaluation. Topics are organized as self-contained modules covering remote sensing and land use classification, digitization of historical maps, and deep learning for prediction. Data science is presented as an input to spatial causal inference techniques: boundary discontinuities, market access approaches, and spatial instrumental variables. Each module pairs a methodological introduction with a recent application from development, trade, or environmental economics. Students will complete hands-on assignments and develop an original policy evaluation proposal that leverages one of these modern measurement tools.