The search for gravitational waves in real life

One of the main challenges faced by GW data-analysis pipelines is to identify rare signals buried in non-Gaussian and non-stationary detector noise. Searches looking for transient signals such as burst or inspiral waveforms are very sensitive to instrumental artifacts and environmental noise, and can therefore be affected by high false alarm rates. Analysis pipelines are designed accordingly to keep the rate of false alarms as low as possible while maintaining thresholds favoring detections. In this lecture I intend to give an overview of the analysis pipeline that has been developed by the LIGO-Virgo Compact Binary Coalescence data-analysis working group. In particular I would like to emphasize the various techniques that have been implemented in the pipeline to minimize the false alarm rate and to discriminate signals from background noise. This includes for instance data quality vetoes and signal-based vetoes, network analysis techniques such as coincidences between detectors, and optimal ranking statistics. Candidates surviving the analysis pipeline are also examined with a detection check-list. An introduction to the software infrastructure that the analysis pipeline relies on will also be given