Portfolio of reproducible data-science projects (forecasting + NLP) on synthetic retail datasets. Notebooks, figures, and READMEs included.

Raster και Raster Stack στην R (workshop FOSSCOMM 2019, Λαμία)

📐 Portfolio project to master statistical modeling, inference, and diagnostics using statsmodels. Includes 15+ concept notebooks, synthetic datasets, custom visual utilities, shared tests with scipy, Streamlit dashboards, Docker support, and markdown cheatsheets.

This notebook explores regularized linear regression from a quantitative and mathematical perspective — deriving cost functions and gradients for OLS, Ridge, Lasso, and ElasticNet, and testing how λ and γ shape the bias–variance trade-off in practice