I study the effects of neighborhood-level inequality on perceived national-level inequality and preferences for redistribution. I construct a novel measure of local inequality using geolocated housing data and elicit perceptions and preferences from an original survey conducted in Barcelona. Local inequality is positively associated with perceived inequality but not with preferences for redistribution. I address endogenous sorting by exploiting quasi-random variation in exposure to new apartment buildings; this increases perceived inequality by 7% and has a positive but not statistically significant effect on demand for redistribution. Effects come from higher perceived income at the top. Local inequality shapes national-level inequality perceptions, but if it influences demand for redistribution, the effect is small. This work suggests that individuals extrapolate from their local environments when forming beliefs and highlights the importance of data granularity when studying neighborhood effects.