Presentations

Predicting fifteen SDG indices for 339 Bolivian municipalities from nighttime lights and AlphaEarth daytime embeddings, and combining both satellite views to monitor local development clusters and spatial development traps.

Invited lecture (in Spanish) on how satellite imagery and spatial data science can measure, predict, and explain uneven development.

Night light seen from space tracks access to infrastructure in Southeast Asia and Japan and shows the gap between regions narrowing. It also reveals how long the 2011 earthquake kept infrastructure in northeastern Japan down.

A framework for using artificial intelligence in learning and research while keeping the work reliable and transparent. It separates producing content from checking it and shows practical ways to combine these tools.

Invited lecture (in Japanese) at the Satellite Data Applications Course, ISEE, Nagoya University.

Seminar at GSID, Nagoya University — using AI (NotebookLM, Google Colab, and GitHub) for data science, econometrics, and research.

A keynote and master class on how artificial intelligence is changing applied econometrics and empirical research. It combines econometric theory with cloud computing, reproducible Python notebooks, and tools powered by artificial intelligence for learning and research.

A workshop on how new data such as satellite images, new methods, and artificial intelligence are changing research on sustainable development in Cambodia and Asia.

A seminar on what satellite images of lights at night can and cannot tell us about local development.

How satellite images, both at night and by day, combined with surveys and maps can reveal development gaps within countries. Together, these sources help predict and understand poverty at many scales.

Presents an easy-to-use Python notebook for processing satellite images of lights at night and using them to study economic activity. It aims to make this kind of data accessible to more regional researchers.

Compares older and newer satellite images of lights at night across 139 countries and over 1,500 regions from 2013 to 2019. The newer images give a more accurate picture of economic inequality within developing countries.

How combining satellite images, household surveys, and machine learning can track regional development in countries where official data are scarce. This helps monitor progress toward the Sustainable Development Goals where surveys are costly or hard to run.

Combines satellite data, a national household survey, and machine learning to map poverty in education, health, and living standards across Cambodia. The maps show where people are most likely to be deprived.

Uses satellite images of lights at night to study how economic activity grows across regions of Java, Indonesia, from 2013 to 2020. Technology and skills spill over between neighboring regions, so education in one place also benefits its neighbors.

Tests whether faster economic growth lowers unemployment across 514 districts of Indonesia from 2011 to 2020. Once the link is measured carefully, the effect appears only in the more industrialized western regions.

Presents an easy-to-use Python notebook for processing satellite images of lights at night and using them to study economic activity. It aims to make this kind of data accessible to more regional researchers.

Compares two sets of satellite images of lights at night to see which better predicts income and inequality across 118 districts of Java from 2013 to 2020. The newer images perform better, especially when comparing places at one point in time.

How income gaps across Turkey’s 81 provinces evolved between 2007 and 2019. The provinces split into six groups, and their neighbors help explain which group each province joins.

An introduction to studying development in fine geographic detail with satellite images and spatial data science. It closes with learning materials and tutorials.

Uses satellite images of lights at night to track inequality across provinces of the Philippines from 2000 to 2020. Inequality fell on average, but the provinces split into two groups that are moving apart.

Shows that Indonesian regions do not all catch up at the same speed, using a method that lets the pace vary from place to place. This gives a more realistic picture of regional gaps across the country’s many islands.

Shows that Indonesian regions do not all catch up at the same speed, using a method that lets the pace vary from place to place. This gives a more realistic picture of regional gaps across the country’s many islands.

Uses satellite images of lights at night to measure income across 274 regions of Southeast Asia between 1998 and 2012. These light-based estimates explain about 60 percent of the differences in official income figures and show how neighboring regions shape each other’s growth.

Uses satellite images of lights at night to compare income gaps across African countries and their regions between 1992 and 2012. The gaps have narrowed, but countries and regions still split into four separate groups.

How gaps in human development across Indonesian districts changed between 2010 and 2018. The gaps narrowed, driven mostly by education, and neighboring districts influence each other’s progress.

How income gaps across Indonesian districts evolved between 2000 and 2017, using new district-level data. Districts form five separate groups, so income gaps persisted even after the country decentralized government.

How income and human development gaps across 151 regions of South America changed between 1990 and 2018. The gaps narrowed overall, and what happens in neighboring regions shapes how fast each region catches up.

How income gaps across Indonesian districts evolved between 2000 and 2017, using new district-level data. Districts form five separate groups, so income gaps persisted even after the country decentralized government.

How productivity differences across 20 Latin American countries evolved between 1980 and 2014. Rather than catching up with each other, the countries split into several separate groups.

Panel discussion at a workshop on cross-disciplinary approaches to good governance, featuring the 5D World Map System.

How gaps in worker productivity, investment in equipment and skills, and efficiency across Indonesian provinces evolved between 1990 and 2010. Provinces did not converge to a single path but formed separate groups.

How productivity differences across 20 Latin American countries evolved between 1980 and 2014. Rather than catching up with each other, the countries split into several separate groups.

Whether Indonesian provinces became more alike in how efficiently they turn resources into output between 1990 and 2010. On average they did, but one type of efficiency splits the provinces into two separate groups.

Whether Indonesian provinces became more alike in how efficiently they turn resources into output between 1990 and 2010. On average they did, but one type of efficiency splits the provinces into two separate groups.

How much economic output each Japanese prefecture produces per unit of carbon emissions, and how this changed between 1992 and 2008. A broader efficiency measure reveals at least two groups of prefectures moving apart, which a simpler measure misses.

How gaps in living standards across Bolivia’s main metropolitan regions changed between 1992 and 2013. The gaps are narrowing, but the least developed regions remain far behind.