As new sources of digital data proliferate in developing economies, there is the exciting possibility that such data could be used to benefit the world’s poor. Recent examples from the research literature show how satellite imagery and deep learning can be used to identify and target pockets of extreme poverty; how mobile phone metadata can help track and stop the spread of Covid-19 and other diseases; how social media analytics can improve disaster response; and how machine learning algorithms can help smallholder farmers optimize planting and harvesting decisions – to name just a few examples.

Through a careful reading of recent research and through hands-on activities and dataset analysis, this course introduces students to the opportunities and challenges for data-intensive approaches to international development. Students should be prepared to dissect, discuss, and replicate academic publications from several fields including development economics, machine learning, information science, and other social sciences. The course requires students to analyze complex and heterogeneous data. Students must have prior graduate training in machine learning, econometrics, or a related field.

Course Logistics

• Professor: Joshua Blumenstock (jblumenstock@berkeley.edu) (pronouns: he/him/his)
• Lectures: Tuesdays, 12:30-3:30, South Hall 210
• Office Hours (for enrolled students): Thursdays, 11:30-12:30, South Hall 203B