This lecture explores Livehoods, a novel approach to understanding urban dynamics by analyzing social media data. By examining check-in patterns, Livehoods map the diverse areas of a city, each telling a unique story. Machine learning is used to study cities based on 18 million check-ins from foursquare. The lecture delves into clustering venues based on pairwise and social distances, building venue graphs, and spectral clustering. It also covers ambiance inference from Airbnb photos, scene recognition, and biases in geo-localized social media, highlighting the urban bias in volunteered geographic information.