Locomotor behaviors result from high dimensional, nonlinear, and dynamically coupled interactions between an organism and its environment. The templates and anchors hypothesis resolves this complexity by anchoring simple mechanical models in physiologically realistic morphologies. By collapsing high-dimensional biological measurements to the simplest representative models, templates have led to insight on neuromechanical control, and enabled the development of bioinspired robots. However, due to the lack of a general methodology for identifying templates from behavioral data, the use of templates to study animal and robot motion has largely been limited to a few well-studied behaviors. Further, reliance upon existing analytic models limits the ability to discover new mechanisms in rich datasets. The promise of big animal mobility datasets from new motion capture labeling methods motivates the aim for a general, data-driven paradigm to identify templates in motion data. We present the vielbein template representation (VTR), a new approach for modeling template-anchor pairs that enables the construction of local model coordinates directly from multidimensional kinematics datasets. We present preliminary results from applying our approach to 3D, whole body ground righting kinematics data collected with Hemidactylus geckos, leading to testable neuromechanical control hypotheses. We conjecture that the VTR approach is a tractable methodology for relating animal and robot data, and will enable new insight into how multifunctional appendages anchor rich behavioral repertoires, and how common physical principles underlie similar maneuvers in morphologically diverse organisms.