MEDIUM_NoteBook

Why we included this project

These notebooks are the working code behind a long-running series of data-science articles, and they read like a hands-on curriculum rather than a loose pile of experiments. Anyone who builds forecasting models day to day will find directly usable material: gradient boosting forecasts, probabilistic forecasting through hitting time, transfer learning for time series, conformal prediction intervals, and practical MLOps patterns, each a self-contained example you can open and run. The collection reaches beyond forecasting too, with clear walkthroughs of model explainability via surrogate models, causal inference through synthetic control, and model selection on imbalanced data. Since every notebook pairs with an explanatory article, it is especially good for learning the reasoning behind a technique instead of just copying a script. For practitioners it also works as a reference of proven implementations to adapt when a similar problem comes up in their own work.