About this site
Built by a practitioner, for practitioners.
Forecast is a practitioner's guide to choosing the right time-series forecasting method. It covers 12 methods with clear trade-offs, a decision worksheet, a comparison matrix, and method-specific pages with Python code and real finance examples.
The audience is working professionals with data and deadlines — financial analysts, data scientists, and engineers who need to forecast revenue, churn, customer adds, demand, or other business time series. Not academics.
The site is written by Shakil Ahmad, a CFA charterholder who uses these methods in production. The tone is precise, practitioner-friendly, and slightly opinionated. No hype.
Three principles behind this site
Practitioners over academics.
Every method page is written for someone with data and a deadline. The question is always: does this work on real data, and how do you know when to use it?
Honest about trade-offs.
No method is universally best. This site tells you when each method wins and when it loses. If a simpler method beats a complex one on your data, use the simpler method.
Ensemble by default.
The default production strategy is not one best method — it is a blend of two or three reasonable methods. M5 first place was an ensemble. The site is built around that reality.