Computational models are chosen considering their adequacy and quality, but the use of dubious data can compromise their outcome. In this article, we confront, in simple terms, a “good” model with “bad” data. The model is a system of two ordinary differential equations (ODEs) of a process of dialysis, and the data are the initial values of the system. For generality, focus is on independence of the application. The model, although analytically soluble, is solved numerically for generality. The 1.st function in the ODEs is the blood urea concentration of a patient under treatment, and the 2.nd is the concentration in the machine (dialyzer). The data supposed uncertain are the initial values to the ODEs, the matter thus being an “initial value problem” (IVP) in ODEs, the data uncertainty being imposed through a Gaussian distribution. The study leads to a simple statistical exploration steered by Monte Carlo simulation, so that an assessment may be made about the extent to which the uncertainty in the data affects the variability of the results of applying the model. The result under observation is the behavior of the final value of the 1.st function, at the final time previously computed without uncertainty. The ODEs and the subsequent Monte Carlo procedure are computed in Python, and an additional contribution is a freely accessible constructed web page, for supplementary reproducibility and verification. There, the user, needing no other software, can both verify our calculations and use other data. Such a web page, as we advocate, offers the confirmation of the computing, and facilitates the desirable academia-industry link.