The implementation of new methods and tools is essential to achieve the desired shift-left concept of vehicle development. The off-car development method outlines a data-driven approach for the continuous adaptation of vehicle software in accordance with the evolving overall system. With the intention of utilising the exploratory nature of road testing in early development stages, off-car robustness testing and automated off-car calibration are integrated into the product development process. These tools serve to bridge the gap between conceptual design and road testing. It is necessary to provide suitable stimuli for simulations and test benches in order to imitate road testing. To enhance the focus on customer applications, this paper uses test drives to emulate in-service operations. Using a big data infrastructure, an algorithm is presented that enables the efficient determination of probabilistic stimuli.