Research in the intersection of economics, machine learning, and spatial data science.

Our project

Nighttime lights (NTL) are satellite observations of light emanating from the Earth's surface at night. NTL data can be a useful proxy for economic activity because they provide consistent measurements worldwide, even where traditional economic statistics are sparse, unreliable, or unavailable. However, the economic interpretation of NTL can vary across cities and urban contexts.

Our project investigates whether systematic biases arising from urban contextual factors—such as land use, population, transportation, and urban form—can be identified and accounted for in the use of NTL as a proxy for economic activity. We use spatial-data and machine-learning methods to learn representations of NTL and its urban context, with a focus on whether those representations improve the measurement of economic activity across heterogenous geographic regions.

Who we are

We are master's students in Harvard Extension School's ALM Data Science and Artificial Intelligence program. The group was formed June 2026 in Cambridge, MA, and our faculty sponsors are from the University of Cambridge, UK. Our dual affiliation makes us Cambridge Squared.