The thesis develops a methodology to improve the analysis of hydrological data by using filtering techniques and mathematical models. In particular, it applies the Kalman filter to datasets from both a water spring and a meteorological station to improve data accuracy and minimize measurement noise. Once the filtering process is completed, the study examines the relationships between meteorological inputs—such as precipitation and temperature—and spring outputs. These correlations are analyzed on a hydrological-year basis, considering the system's response times to external influences. The results intend to assess the reliability of meteorological data from distant stations and their effectiveness in predicting future spring behavior.