The viral spread of the social web is quickly becoming a rich source of information for recommendation systems and other personalised services. Semantics and web science help to decipher and render the chaotic and disconnected data on the social web, and turn it into clean, structured, and well integrated knowledge. In this talk I will briefly argue for the need for semantics and web science to tame the social web, then I will describe our work and results with integrating semantics, the social web, and the real-world in the context of the Live Social Semantics application. Finally I will highlight a number of recommendation services that you could help building for our Live Social Semantics platform.