The interpretation of stabilization diagrams and the tracking of modes over time are classical tasks in operational modal analysis. In this work, we present a method based on a greedy clustering that efficiently extracts the modal parameters from stabilization diagrams with covariance-driven subspace identification and integrated uncertainty quantification. Stability criteria are strongly based on the estimated modal parameter uncertainties. From the analysis of one, or several, stabilization diagrams, a set of modal parameters is defined, for it to be tracked over time in the next analysis step. The tracking is performed by an active search for the reference parameters in new datasets by combining stability criteria with efficient search heuristics. The resulting algorithm is efficient in tracking large parameter changes, which is demonstrated on the S101 Bridge benchmark under artificially introduced damages, as well as on data of the Munich Test Bridge, where the modal parameters are strongly affected by temperature variations.