As building systems grow taller, more lightweight and efficient, they often become increasingly sensitive to the effects of wind. In such situations, habitability limit states govern their design, as wind-induced accelerations increase and become more perceptible to occupants with the potential to adversely affect occupant comfort. In particular, since this limit state involves human perceptions, it can be quite challenging to accurately quantify the level of acceleration that would be acceptable, leading to the lack of a unified standard for design. As the dynamic responses that must ultimately be compared to this ambiguous limit state are characterized by mass, stiffness, and damping, accurate prediction of these parameters also becomes increasingly critical. While mass and stiffness are assumedly readily determined in the design stage, damping continues to elude structural engineers, who remain reliant on rudimentary estimates that are largely based on the building's primary material: steel or reinforced concrete. This often proves problematic as damping is a particularly critical parameter in the habitability design of flexible structures. In fact, full-scale monitoring efforts around the world have shown that many tall buildings exceed accelerations predicted in the design stage and that in-situ damping values are often lower than assumed. This is further compounded when in-situ frequencies are found to disagree with the FE model predictions, which can further contribute to habitability issues. Estimating and understanding these dynamic properties is further complicated in the presence of amplitude dependence and complex building behaviors such as coupling.  This research addresses the uncertainties associated with the habitability design of tall buildings by viewing the unique insights afforded by full-scale monitoring. This effort begins by offering a pseudo-full-scale evaluation of occupant comfort to better quantify habitability performance under lateral and torsional responses. These full-sale responses are then viewed through the lens of structural system behavior, i.e., the degree of cantilever action displayed by the system, to provide designers with a set of heuristic guidelines to inform a more accurate prediction of the periods of tall buildings in the design stage. By then introducing a new wavelet-based system identification framework, large amplitude full-scale responses are mined to gain greater insight into the level of energy dissipation at critical design limit states that then drives a more robust and effective predictive model for inherent damping based on this system behavior descriptor. The end result of this dissertation is a suite of guidelines, frameworks and models that enables a more accurate prediction and evaluation of habitability performance of tall buildings.