Applied Time Series Forecasting

About the Lab

Building reproducible forecasting pipelines that connect research, open data, and practical implementation.

Mission

Sentinel Forecasting Lab is an independent initiative dedicated to time series forecasting and machine learning. Rather than presenting isolated models, the tutorials explain complete workflows, from data acquisition and feature engineering to model training, evaluation, and interpretation.

Philosophy

Reproducibility Transparency Educational value

Whenever possible, every tutorial relies on publicly available datasets and open-source software so that readers can reproduce, adapt, and extend the results on their own systems.

Topics currently explored

Hydrology Energy Time Series Forecasting Machine Learning ERA5 Open Data

Current tutorials cover applications in hydrology, energy forecasting, meteorology, and multivariate time series analysis, with particular emphasis on interpretable models and the integration of heterogeneous data sources.

About the author

Sentinel Forecasting Lab is an independent initiative created and maintained by Eric Duhamel. Its ambition is to make advanced forecasting techniques accessible through rigorous yet practical tutorials that bridge the gap between scientific research and operational applications.

Whether you are a student, researcher, or practitioner, the objective is to provide resources that can be directly reproduced, adapted, and extended to your own forecasting problems.