This project investigates the spatial relationship between race, education attainment, and unemployment in Montevideo, Uruguay, testing whether geographic context improves predictive power over traditional linear models. As the sole analyst, I designed a comparative modeling workflow using ArcGIS, training both an Ordinary Least Squares (OLS) regression and a Multiscale Geographically Weighted Regression (MGWR) to predict unemployment rates across census tracts while incorporating spatial autocorrelation and racial demographic data. The final deliverable demonstrates my ability to work across the full lifecycle of a spatial data science project, from data acquisition and exploratory spatial analysis to advanced geostatistical modeling and interpretation of coefficients that became significant only when geography was accounted for. The MGWR model explained 90% of the variance in unemployment rates, a substantial improvement over OLS, revealing that the effects of education and race on unemployment vary significantly across Montevideo—with the highly educated, predominantly White neighborhoods in the south showing markedly different patterns than the less White, lower-education areas at the city's edges.