My choice of research projects over the past decade has been rather eclectic, covering aspects relating to both seismic sources and Earth structure. There is, however, a consistent theme, and that is a fascination with the large variety of methods for extracting pertinent information from seismic data. In this thesis, I give an brief, largely chronological outline of the steps and insights that have informed my current thinking on this theme, going into more detail on two methods that I have recently developed: FLEXWIN, for automatically selecting the most appropriate time-windows on complex seismograms in which to make measurements, and WaveLoc, for automatically detecting and locating seismic phenomena from continuous waveform data by exploiting the coherence of information across a seismic network. Such tools, based on the the integration and automation of practical seismological "know-how", allow us to exploit seismological data more completely, and are becoming indispensable in the context of the enormous volume of data produced by modern seismic networks.
