TY - JOUR TI - Faces in places DO - https://doi.org/doi:10.7282/T3GM89K1 PY - 2016 AB - In this study, we explore a novel approach to measure fine-grained (photo-level) diversity using Instagram. We compare and contrast these new measures of diversity with traditional metrics (i.e. census). We discuss the merits and shortcomings of supplementing traditional census figures with these new measures. Further, we explore the predictive capacity that this new metric has over socio economic outcomes, namely income inequality. We find that using our ne-grained metric for measuring diversity in interactions produces very different results compared to traditional census measures. We also determine that diversity (specifically photo based entropy in age and race) are associated with income inequality and the combined model is significantly (though weakly) predictive of inequality. Neighborhoods that have high scores in racial diversity seem to have a correlation with lower inequality, while neighborhoods that have high scores in age diversity seem to have a correlation with higher inequality. We discuss the possible implications of this work on research in sociology and associated areas and suggest further work based on these findings. KW - Electrical and Computer Engineering KW - Multiculturalism KW - Social media LA - eng ER -