The large flashes of radiation that are observed in GRBs are generally believed to arise in a relativistic jetted outflow. This thesis addresses the question of how and where in the jet this radiation is produced. It further explores the jet properties that can be inferred from the observations made by the Fermi GST that regularly observes GRBs in the range 8 keV - 300 GeV.  In my analysis I focus on the observational effects of the emission from the jet photosphere. I show that the photosphere has an important role in shaping the observed radiation spectrum and that its manifestations can significantly vary between bursts. For bursts in which the photospheric  emission component can be identified, the dynamics of the flow can be explored by determining the  jet Lorentz factor and the position of the jet nozzle. I also develop the theory of how to derive the properties of the outflow for general cases. The spectral analysis of the strong burst GRB110721A reveals a two-peaked spectrum, with the peaks evolving differently. I conclude that three main flow quantities can describe the observed spectral behaviour in bursts:  the luminosity, the Lorentz factor, and the nozzle radius. While the photosphere can appear like a pure blackbody it can also be substantially broadened, due to dissipation of the jet energy below the photosphere. I show that Comptonisation of the blackbody can shape the observed spectra and describe its evolution. In particular this model can very well explain GRB110920A which has two prominent breaks in its spectra.  Alternative models including synchrotron emission leads to severe physical constraints, such as the need for very high electron Lorentz factors, which are not expected in internal shocks. Even though different manifestations of the photospheric emission can explain the data, and lead to ambiguous interpretations, I argue that dissipation below the photosphere is the most important process in shaping the observed spectral shapes and evolutions.