Currently, experimental and field data are main sources of estimating uncertainties. When adequate data are not available, engineering and intuitive judgment have been used, which without any rationale, may result in further propagation of error in estimating uncertainties. This paper describes the method of data mining for uncertainty estimation through experts. The method is flexible and can be crafted for a specific application in estimating uncertainties. The data collection is performed in multiple rounds to minimize the bias in responses. Also, the experts can be provided with ample information to refine the process of data mining and bias management. They may be provided with a scenario that replicates, or presents a digital twin of a known case, to bring their thoughts to focus on a specific parameter of concern. Modes of bias relate to an expert’s perception of the problem, their cognitive response in approaching it, and reaction to similar events. Modes of bias are reviewed and strategies to mitigate their effects are presented. This method of uncertainty estimation is useful when the nature of a parameter does not lend itself to conventional methods of data compilation. Thus, the expert opinion can alternatively be used to mine data and estimate the uncertainty. The method is especially applicable when the geometry of a system does not allow tests through experimentation; and field measurements do not provide meaningful data because of low levels of mechanical effects. A procedure for conducting the method is presented along with a discussion on potential challenges.