Data analysis plays two major roles in gravitational wave astronomy. First, based on information from astrophysicists and source modelers, the data analyst tries to determine whether a signal of a given type exists in the data stream being produced by the detector. Second, when a signal is found, the data analysts tries to extract the signal waveform, thus providing feedback to the source modeler and the astrophysicist as to what gravitational wave sources exist. In some cases, both of these tasks can be accomplished simultaneously, although in general one uses a "quick and dirty" method to search all the data for the existence of a signal and then a more refined method to nail down the exact waveform once it is located.

For LISA, the primary data analysis technique for both stages of the analysis is expected to be the standard matched filtering algorithm, which is already understood and implemented in the context of ground-based interferometers. Matched filtering, in essence, correlates the data at each moment with a bank of theoretically predicted waveforms, and claims a detection is made when a statistically significant correlation occurs. It will therefore be necessary to design template banks for LISA sources, many of which are not shared with ground-based interferometers. Also, in some cases where there is an overlap in the source population (such as with comparable-mass black hole mergers, for which the waveforms simply scale with mass), it is unclear whether source modeling will have reached a point where templated searches are feasible by the time the search has begun. Therefore, it is likely - and in the case of signals from unpredicted or unmodeled sources (such as stochastic sources), certain - that techniques other than matched filtering will be required as well.

Since there is only a single LISA, the ground-based searches relying on coincident detection in multiple instruments will need to be modified. Also, triangulation of sources using three or more detectors will be infeasible, but for long-lasting sources positioning using Doppler modulations will be possible. Clearly, then, while data analysis for ground-based observatories will provide a base from which to build, there is a great deal of new ground to be broken for LISA data analysis. Initially, the CGWA will focus on two LISA sources for data analysis, although we can expect that the focus may broaden as the center matures.