Waterfall diagrams, W`, where each sheet has M × N dimension. M corresponds to the number of divisions (number of rows), N is the number given by the trial period P_ℓ (i.e., P_t of the iteration ℓ) divided by a chosen time bin ∆t: N = round [P_ℓ/∆t]. A data point has an amplitude given by the number of counts. L corresponds to the total number of trial periods P_ℓ or number of iterations. For this particular case, L = 3, M = 5, and N = 9. The PCA is computed for every , which returns (ℓ, m) eigenvalues and eigenvectors.