please write the research paper on this Predicting Financial Inclusion Gaps Using Alternative Data (Mobile & Satellite) for Underbanked Populations Why this works well for a scholarship committee: Clear social impact — over 1.4 billion adults globally remain unbanked; this ties directly into UN SDG 8 and 10, which most scholarship boards (Fulbright, Chevening, DAAD, Erasmus Mundus, etc.) explicitly favor. Strong stats/math fit — you'd use techniques you already have a foundation for: logistic regression, hierarchical/multilevel models, spatial statistics, and gradient boosting, plus causal inference methods (propensity score matching, difference-in-differences) if you want to go further. Feasible with public data — no need for proprietary bank records to get started. Core research question: Can mobile phone usage patterns and satellite-derived socioeconomic indicators predict financial inclusion gaps at a sub-national level, and which features matter most for targeting interventions? Data sources (all public/accessible): World Bank Global Findex Database (financial inclusion survey data) DHS (Demographic and Health Surveys) — household wealth indices Satellite night-light data (NOAA VIIRS) as a proxy for regional economic activity Mobile money penetration data (GSMA reports) Methodology outline: Merge geospatial + survey data at district/regional level Build predictive models (start simple: logistic regression baseline → random forest/XGBoost) for financial inclusion status Use SHAP values or similar for interpretability — important since reviewers like policy-relevant, explainable outputs Optionally, add a causal layer: does mobile money rollout cause increased inclusion (using a natural experiment or DiD design)?