The Data Revolution
This year’s Griffin Applied Economics Incubator explores the potential implications of using machine learning and big data to understand economic questions.
For years, firms have achieved incredible successes in harnessing the power of bigger and better datasets using novel techniques from machine learning, computer science, and statistics. This revolution has helped make substantial progress on long-standing problems, such as speech and image recognition.
Although many questions in economics are fundamentally distinct from these applications, they could conceivably benefit from the big data and machine learning revolution. Examples include assessing causal effects of policy interventions, modeling new types of economic data such as text or networks, and targeting interventions to specific groups or markets.
Can off-the-shelf methods be used to answer such economic questions, or do they need to be modified in fundamental ways? And what type of theoretical guarantees do these methods possess? More broadly, can the knowledge from disciplines outside of economics be used to exploit the power of big data, and be successfully applied to areas of economic research?
The Incubator will bring together an interdisciplinary team of researchers and thought leaders to unravel some of the methodological and substantive questions which arise in the age of big data.