Monday, November 22nd, 2021
11:00 a.m. – 12:15 p.m. EST
via Zoom
Poverty statistics are conventionally compiled using data from household income and expenditure survey or living standards survey. This study examines an alternative approach in estimating poverty by investigating whether readily available geospatial data can accurately predict the spatial distribution of poverty in Thailand. In particular, geospatial data examined in this study include night light intensity, land cover, vegetation index, land surface temperature, built-up areas, and points of interest. The study also compares the predictive performance of various econometric and machine learning methods such as generalized least squares, neural network, random forest, and support vector regression. Results suggest that intensity of night lights and other variables that approximate population density are highly associated with the proportion of an area’s population who are living in poverty. The random forest technique yielded the highest level of prediction accuracy among the methods considered in this study, perhaps due to its capability to fit complex association structures even with small and medium-sized datasets. Moving forward, additional studies are needed to investigate whether the relationships observed here remain stable over time, and therefore, may be used to approximate the prevalence of poverty for years when household surveys on income and expenditures are not conducted, but data on geospatial correlates of poverty are available.
About the Speakers:
Neeti Pokhriyal is a visiting scholar in the Department of Computer Science, Dartmouth College, where she was employed as a postdoc from 2019-2021 funded by the Institute for Security, Technology and Society. She is interested in modeling scenarios characterized by noisy, uncertain, and high-dimensional data coming from heterogeneous sources, with emphasis on reasoning under uncertainty and quantifying biases. She seeks understanding of problems targeting sustainable human development using knowledge inspired and data-driven computational techniques and is interested in exploring evidence-driven policy planning.
She was awarded a seed grant from the Arthur L. Irving Institute for Energy and Society, Dartmouth College in 2020 to propose methods that enable frequent evaluations of energy deficit in poorer economies in absence of any official surveys. She has collaborated with the Inter-American Development Bank, DC on studying poverty and inequality for Haiti from satellite imagery and mobile phone data.
Her doctoral work was awarded the Chih Foundation Research Award in 2019, which is a single award of USD 2.5K given for innovative research for the betterment of society at University at Buffalo, State University of New York, from where she completed her Ph.D in Computer Science at the Center for Unified Biometrics. During her Ph.D, she led a project funded by Gates Foundation for mapping multi-dimensional poverty using mobile data and has teamed with the National Statistics Office of Senegal and Sonatel. She has also collaborated with the Oversees Development Institute (ODI), London, and Datapop Alliance regarding poverty mapping work in Senegal.
Prior to Ph.D, she was a researcher in the Computer Science and Mathematics Division, Oak Ridge National Laboratory, and obtained here Masters in Computer Science from University of California, Riverside, where she received the Dean’s Distinguished Fellowship. She also has an undergraduate degree in Computer Science and Engineering with honors.
Nattapong Puttanapong is a professor and economist in profession. He is presently an Assistant Professor at Thammasat University and the senior economist at the research institute of Thailand’s Ministry of Finance. He has also worked as a consultant to various government agencies and international organizations such as OECD, ILO, World Bank, ADB and JICA. He was awarded with the Royal Thai Government Scholarship, through which he obtained his Ph.D. in Regional Economics from Cornell University. Dr. Puttanapong’s research interests are in the areas of economic modelling, spatial econometrics, and socioeconomic disparities.
Damien Jacques is the lead data scientist of Rubyx, a company designing risk and profit optimization solutions for banking institutions in emerging countries. He has led projects across the globe requiring (i) designing and implementing algorithms to extract key insights from large unstructured data, (ii) develop strategies to leverage the entire data value chain of companies and development agencies; and (iii) successfully scale up data solutions in complex and multi-stakeholder ecosystem. Damien has extensive expertise in the use of non-traditional data for poverty monitoring and has contributed to the following projects:
– Combining cell phone and satellite image data for improved multidimensional poverty monitoring in Senegal. The results have been published in PNAS (University of Buffalo, Orange).
– Tracking a socio-economic crisis in central America and its impact on poverty using a series of indicators generated from the mobile phone activity of users in the country before and after the crisis (Inter-American Development Bank).
– Estimate poverty at the individual level using mobile phone data in Uganda in order to target the poorest with cash transfers (Dalberg Data Insights, GiveDirectly).
– He also co-organized the poverty mapping initiative with the World Bank and the Qatar Computing Institute.The goal of the workshop was to share knowledge and make plans to better combine traditional development data (such as household surveys, labor force surveys and censuses) with complementary sources of big data (satellite, mobile phones, social media) towards achieving more accurate, timely and cost-effective measures of poverty. See blog Making a better Poverty Map.
Damien holds a Ph.D. in Bioengineering and Agronomy from the catholic University of Louvain.
