Abstract
Isotopic composition modelling is a key aspect in many environmental studies. This work presents Isocompy, an open source Python library that estimates isotopic compositions through machine learning algorithms with user-defined variables. Isocompy includes dataset preprocessing, outlier detection, statistical analysis, feature selection, model validation and calibration and postprocessing. This tool has the flexibility to operate with discontinuous inputs in time and space. The automatic decision-making procedures are knitted in different stages of the algorithm, although it is possible to manually complete each step. The extensive output reports, figures and maps generated by Isocompy facilitate the comprehension of stable water isotope studies. The functionality of Isocompy is demonstrated with an application example involving the meteorological features and isotopic composition of precipitation in N Chile, which are compared with the results produced in previous studies. In essence, Isocompy offers an open source foundation for isotopic studies that ensures reproducible research in environmental fields.