Template for forecasting data science project and identify consumption profiles in time series
-
Updated
Oct 3, 2023 - Jupyter Notebook
Template for forecasting data science project and identify consumption profiles in time series
A production-ready, end-to-end system for spare parts inventory management. This project combines advanced forecasting (TSB, Prophet, SARIMAX), mathematical optimization (CPLEX), and supply chain simulation (anyLogistix) to achieve 98.9% forecast accuracy and a 39% ROI improvement.
Benchmarking time-series foundation models (Chronos-Bolt, zero-shot) vs. supervised (PatchTST) and classical (seasonal-naive, Croston) baselines on the M5 Walmart dataset, scored with MASE and WQL. No single model dominates: foundation/deep models win on dense SKUs, classical methods win on the intermittent tail.
End to end SKU level sales forecasting system built for a tire distributor. LightGBM global model, Croston SBA for intermittent demand, MinT hierarchical reconciliation, and A to D confidence grading across 1,368 products.
Supply chain planning for small manufacturers: time-phased MRP, intermittent-demand forecasting, rough-cut capacity. Benchmarked against Excel and ERP min/max on 3,593 real demand series (two public datasets, 90-day holdout) — reproducible method, including where it loses. Runs offline.
Two-week customer x product demand forecasting for intermittent FMCG sales (global LightGBM/Tweedie, CSV/PostgreSQL, Excel export)
Demand-regime instability detection for daily SKU demand series with STABLE / DRIFT / SHIFT state outputs.
Classifies demand by pattern, routes each pattern to the method built for it, and evaluates by walking forward in time.
Procurement-oriented demand forecasting for many-SKU intermittent demand: quantile-of-sum ordering, tiered service levels, order-level backtesting. Demonstrated on M5.
To associate your repository with the intermittent-demand topic, visit your repo's landing page and select "manage topics."