Previous research has demonstrated the potential benefits of thermal aware load placement and thermal mapping in cool-intensive environments such as data centers. However, it has proved difficult to apply existing techniques to live data centers because of models that are either unrealistic, require extensive sensing instrumentation, or because their creation is disruptive to the data center services. The work presented in this paper discusses techniques and their associated challenges with respect to creating an adaptive and non-invasive method of creating realistic and low-complexity thermal models using built-in and ambient sensors. Uses of these techniques can vary from assessing the thermal efficiency of the data center to designing a thermal-aware job scheduler to lower total cost of ownership (TCO). Specifically, this paper proposes: i) a noninvasive thermal modeling software architecture that uses onboard, ambient and software sensors ii) and four different ways of leveraging the gathered data from an experimental application of the architecture to improve the greenness of the data center and our understanding of the thermal behavior of a data center.