Future Work:
Enhancing Poverty Measurement

Traditional poverty measurement based on surveys does not provide granular data on income/consumption, wealth, or development at many specific locations of interest, thus limiting effective interventions targeted at communities. It is, therefore, critical to develop surrogate predictive methods to estimate georeferenced poverty measures at any location in a region using widely available data such as social media, crowd-sourced maps, or remote sensing images. To address this critical gap, my research aims to introduce an integrated geostatistical machine learning framework that efficiently imputes missingness and quantifies uncertainty at the data processing stage; spatial cross-validation and debiasing techniques at the inference stage; and robust resource optimization under uncertainty at the policy implementation stage. 
..................................................................................................................................................................................