With the aim to help users navigate through the large list of channels and programs now available, we model mood from television content, using close-captions provided by an industry partner. We mine this data using a tridimensional model of affect to represent mood. With our approach, we can model the mood of a particular program, or even the topics mentioned in a program (how does this show portrays a certain politician, city, or a company?), both using historical data and real-time data.