B-AMA: a new Python protocol for hydrological predictions using data-driven models

The objective of this interactive poster session is to show the main features of B-AMA (Basic dAta-driven Models for All), an easy, flexible, fully coded Python-written protocol for the application of data-driven models (DDM) in hydrology. The protocol is specifically tailored for early career scientists with little background in coding, to foster them through the development of DDMs for hydrological forecasting while ensuring that none of the fundamental methodological steps is overlooked.

While a Jupyter notebook is already available online to guide the users through the protocol employment, during the session the interested audience can learn the main features of the software (data splitting, feature selection, hyperparameter optimization, and performance metrics) by running several practical hydrological workflows. The session will couple the visual representation B-AMA’s methodology with some laptop-based experiments, including rainfall-runoff, hydropower, and groundwater forecasts. We also allow loading customized csv data to deliver a first-hand experience of the protocol forecasting ability on the user’s specific case study, thanks also to the embedded visualization tools, which facilitate the efficient investigation and communication of results.