FPP3 Forecasting Explorer — Supply Chain Demand

Compare forecasting methods on supply chain demand data. Default scenario: electronics distributor with Q4 peaks. Adjust patterns, toggle models, and see which method wins. All calculations match fpp3 formulas.

Demand Pattern

Forecasting Models

Chart View

—
Best Model
—
vs Naive Improvement
—
Forecast Horizon
—
Models Compared

Advisory Panel

Excel vs fpp3 — Effort Comparison

Reference: Hyndman, R.J. & Athanasopoulos, G. (2021) Forecasting: Principles and Practice, 3rd ed., OTexts. Models: Naive (NAIVE), Seasonal Naive (SNAIVE), SES (ETS(A,N,N)), Holt's Linear (ETS(A,A,N)), Holt-Winters Additive (ETS(A,A,A)). Smoothing defaults: α=0.3, β=0.1, γ=0.3. MASE uses seasonal naive (m=12) as scaling denominator, consistent with fpp3 accuracy(). Decomposition is classical additive (12-period centered MA) — the blog post uses STL, so exact values differ. Default scenario: electronics distributor, 48 months, Q4 peaks.