Método Tornado como marco gráfico exploratorio para distribuciones discretas Con numero primos

This manuscript introduces the Tornado Method as a computational and graphical framework for visual inspection, exploratory data analysis, and anomaly detection in discrete frequency distributions. The methodology formalizes rules for transforming numerical datasets using digit sums and digital roots, representing exact counts for each target through ellipses whose areas are strictly proportional to the observed frequencies. As a numerical benchmark and case study, sequences of prime numbers up to one million are analyzed, documenting exact arithmetic constraints and distribution convergence. Furthermore, a statistical control scheme is integrated using total variation distance ($D_{TV}$) and Monte Carlo simulations to calibrate tolerances, evaluate sensitivity to data perturbations, establish negative controls, and define the method's limits of identificability.