The Chandra X-ray Observatory has been producing significant astronomical discoveries since its launch by NASA in July of 1999. Now in Chandra's second decade of science, the Chandra operations team is using Python to create predictive thermal models of key spacecraft components. These models are part of the mission planning and command load review process to ensure that the series of planned observations and attitudes for each week will maintain a safe thermal environment. This poster describes the Xija modeling framework that is used to create, calibrate, and compute Chandra thermal models. This package provides a generalized framework to model complex time series data using a network of connected nodes with pluggable model components that define the node interactions. At present the model components include thermal conduction and passive and active heating elements, but the framework itself is general and could be used for other applications. A key feature is a GUI fitting tool that allows for rapid evaluation of model fit results and interactive many-parameter fits of large time-series datasets using the Sherpa fitting package.