The detection of GW150914 has marked the start of the era of gravitational wave (GW) astronomy, and the number of detections available for astrophysical interpretation has been growing ever since. Similarly, this thesis progresses from considering individual signals to analysing the population of gravitational wave sources.

We explore how much we expect to be able to learn from a loud binary black hole merger like GW150914, were it to appear in the detector networks we currently have, as well as those which are being planned or under construction. Afterwards, we develop a model which allows us to obtain unbiased estimates of the rate with which the universe produces such signals, and the distribution of intrinsic gravitational wave source parameters such as the component masses, despite the fact that the observed distribution is inherently biased by the fact that source parameters affect the loudness of the emitted signal, and contaminated by terrestrial noise artefacts inherent to the detectors. We apply this method to a number of test cases.